Snakemake workflow: Meta-analysis of GWAS of gestational duration, preterm and post-term deliveries (EGG Consortium)

public public 1yr ago Version: v1.0.0 0 bookmarks
View Workflow
Report
metagwas
metagwas
View Workflow

Snakemake workflow: Meta-analysis of GWAS of gestational duration, preterm and post-term deliveries (EGG Consortium)

This repository contains code for Genetic effects on the timing of parturition and links to fetal birth weight .

Citation

Contact

  • Pol Sole-Navais (@psnavais)

  • pol.sole.navais@gu.se

Code Snippets

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]

cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tlocus\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tlocus\n', file= snakemake@output[[2]])


prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6

d= fread(snakemake@input[[1]], select= c('ID', 'CHR', 'POS', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue', 'EAF'))
d$MAF= ifelse(d$EAF>0.5, 1 - d$EAF, d$EAF)

x= fread(snakemake@input[[2]], select= c('ID', 'N','BETA', 'SE', 'pvalue', 'EAF'))

x$MAF= ifelse(x$EAF>0.5, 1- x$EAF, x$EAF)

names(x)= c('ID', 'N', 'beta', 'se', 'p', 'eaf', 'maf')

d= inner_join(d, x, by= 'ID')

if (sum(is.na(d$eaf)) == nrow(d)) {
d$maf= d$MAF
} 

z= fread(snakemake@input[[3]])

z$CHR= as.numeric(gsub('chr', '', z$chr))

z$locus= 1:nrow(z)


funk= function(i) {
        row= z[i,]
	locus= paste0('locus_', i)
        temp_df= filter(d, CHR== as.integer(row[, 'CHR']), POS >= as.integer(row[, 'start']), POS<= as.integer(row[, 'stop']))

	if (nrow(temp_df)== 0) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0,  PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
	pph_list[[i]]= PPH
	res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')

	} else {
	temp_df= filter(temp_df, SE>0, se>0)
	if (grepl('PCOS|miscarriage|POP|endometriosis|Preeclampsia|leiomyoma_uterus', snakemake@input[[2]])) {
        if (grepl('PCOS', snakemake@input[[2]])) {s_pheno=  (1184 + 670 + 157 +658 +984 + 485 + 462 )/ (1184 + 670 + 157 +658 +984 + 485 + 462 + 5799 + 1379 +2807 +6774 +2963+ 407 + 96172)}
        if (grepl('miscarriage', snakemake@input[[2]])) {s_pheno=   49996 / ( 174109 + 49996)}
        if (grepl('POP', snakemake@input[[2]])) {s_pheno= 7053 / (57407 + 7053) }
        if (grepl('endometriosis', snakemake@input[[2]])) {s_pheno= 1496 / (192678 + 1496 )}
        if (grepl('Preeclampsia', snakemake@input[[2]])){ s_pheno= 4630/ (4630 + 373345)}
        if (grepl('leiomyoma_uterus', snakemake@input[[2]])){ s_pheno= ( 14569) / (85792 + 14569)}
        if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.067)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.122)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

        data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'cc', snp= temp_df$ID, s= s_pheno, MAF= temp_df$maf)

        } else { 
	if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.067)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.122)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

        data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'quant', snp= temp_df$ID, MAF= temp_df$maf)
}
myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
	if (length(myres)==1 ) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
        pph_list[[i]]= PPH
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
	next 
	} else {
	PPH= data.frame(t(myres[[1]]))
        PPH$locus= locus
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$locus= locus
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)

}
}
}

mclapply(1:nrow(z), funk, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of colocalization/coloc_BW_GA_GW.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]

cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tlocus\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tlocus\n', file= snakemake@output[[2]])


prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6

d= fread(snakemake@input[[1]], select= c('ID', 'CHR', 'POS', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue', 'EAF'))
d$MAF= ifelse(d$EAF>0.5, 1 - d$EAF, d$EAF)

x= fread(snakemake@input[[2]], select= c('ID', 'TOTALSAMPLESIZE','BETA', 'SE', 'pvalue', 'EAF'))

x$MAF= ifelse(x$EAF>0.5, 1- x$EAF, x$EAF)

names(x)= c('ID', 'N', 'beta', 'se', 'p', 'eaf', 'maf')

d= inner_join(d, x, by= 'ID')

if (sum(is.na(d$eaf)) == nrow(d)) {
d$maf= d$MAF
} 

z= fread(snakemake@input[[3]])

z$CHR= as.numeric(gsub('chr', '', z$chr))

z$locus= 1:nrow(z)


funk= function(i) {
        row= z[i,]
	locus= paste0('locus_', i)
        temp_df= filter(d, CHR== as.integer(row[, 'CHR']), POS >= as.integer(row[, 'start']), POS<= as.integer(row[, 'stop']))

	if (nrow(temp_df)== 0) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0,  PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
	pph_list[[i]]= PPH
	res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')

	} else {
	temp_df= filter(temp_df, SE>0, se>0)
	s_pheno= 0.067
	data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF)
	data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'cc', snp= temp_df$ID, s= s_pheno, MAF= temp_df$maf)
	myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
	if (length(myres)==1 ) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
        pph_list[[i]]= PPH
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
	next 
	} else {
	PPH= data.frame(t(myres[[1]]))
        PPH$locus= locus
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$locus= locus
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)

}
}
}

mclapply(1:nrow(z), funk, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of colocalization/coloc_GA_vs_PTD_GW.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
library(data.table)
library(dplyr)
library(coloc)

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6

d= fread(snakemake@input[[1]])
d= select(d, ID, CHR, POS, TOTALSAMPLESIZE, BETA, SE, pvalue, EAF)
d$MAF= ifelse(d$EAF>0.5, 1 - d$EAF, d$EAF)

x= fread(snakemake@input[[2]])

x$BETA= ifelse(x$REF > x$EFF, -1 * x$BETA, x$BETA)

x= select(x, ID, TOTALSAMPLESIZE, BETA, SE, pvalue, EAF)
x$MAF= ifelse(x$EAF>0.5, 1- x$EAF, x$EAF)
names(x)= c('ID', 'N', 'beta', 'se', 'p', 'eaf', 'maf')

d= inner_join(d, x, by= 'ID')

if (sum(is.na(d$eaf)) == nrow(d)) {
d$maf= d$MAF
} 

z= fread(snakemake@input[[3]])
z$CHR= ifelse(z$CHR== 'X', '23', z$CHR)
z$CHR= as.integer(z$CHR)

z1= fread(snakemake@input[[4]])
z1$CHR= ifelse(z1$CHR== 'X', '23', z1$CHR)
z1$CHR= as.integer(z1$CHR)
z1= filter(z1, nearestGene== 'LRP5' | nearestGene== 'SCML4')

z= rbind(z, z1)

pph_list= list()
res_list= list()

for(i in 1:nrow(z)) {
        row <- z[i,]
	locus= paste0('chr', row[,'CHR'], '_', row[,'nearestGene'])
        temp_df= filter(d, CHR== as.integer(row[, 'CHR']), POS >= as.integer(row[, 'pos1']), POS<= as.integer(row[, 'pos2']))

	if (nrow(temp_df)== 0) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0,  PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
	pph_list[[i]]= PPH
	res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res

	} else {
	temp_df= filter(temp_df, SE>0, se>0)
	s_pheno= 0.067
	data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF)
	data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'cc', snp= temp_df$ID, s= s_pheno, MAF= temp_df$maf)
	myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
	if (length(myres)==1 ) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
        pph_list[[i]]= PPH
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	next 
	}
	PPH= data.frame(t(myres[[1]]))
	PPH$locus= locus        
	pph_list[[i]]= PPH
        res= myres[[2]]
        res$locus= locus
	res_list[[i]]= res
}
}

pph= data.frame(do.call('rbind', pph_list))
res= data.frame(do.call('rbind', res_list))


write.table(pph, snakemake@output[[1]], sep= '\t', row.names=F, col.names= T, quote=F)
write.table(res, snakemake@output[[2]], sep= '\t', row.names=F, col.names= T, quote=F)
R dplyr data.table Quant coloc From line 1 of colocalization/coloc_GA_vs_PTD.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
library(data.table)
library(dplyr)
library(coloc)

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6

d= fread(snakemake@input[[1]], select= c('ID', 'CHR', 'POS', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue', 'EAF'))
d= select(d, ID, CHR, POS, TOTALSAMPLESIZE, BETA, SE, pvalue, EAF)
d$MAF= ifelse(d$EAF>0.5, 1 - d$EAF, d$EAF)

x= fread(snakemake@input[[2]], select= c('ID', 'REF', 'EFF', 'BETA', 'EAF', 'SE', 'N', 'pvalue'))

x$BETA= ifelse(x$REF > x$EFF, -1 * x$BETA, x$BETA)

x= select(x, ID, N, BETA, SE, pvalue, EAF)
x$MAF= ifelse(x$EAF>0.5, 1- x$EAF, x$EAF)
names(x)= c('ID', 'N', 'beta', 'se', 'p', 'eaf', 'maf')

d= inner_join(d, x, by= 'ID')

if (sum(is.na(d$eaf)) == nrow(d)) {
d$maf= d$MAF
} 

z= fread(snakemake@input[[3]])
z$CHR= ifelse(z$CHR== 'X', '23', z$CHR)
z$CHR= as.integer(z$CHR)

pph_list= list()
res_list= list()

for(i in 1:nrow(z)) {
        row <- z[i,]
	locus= paste0('chr', row[,'CHR'], '_', row[,'nearestGene'])
        temp_df= filter(d, CHR== as.integer(row[, 'CHR']), POS >= as.integer(row[, 'POS']) - 250000, POS<= as.integer(row[, 'POS']) + 25000)

	if (nrow(temp_df)== 0) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0,  PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
	pph_list[[i]]= PPH
	res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res

	} else {
	temp_df= filter(temp_df, SE>0, se>0)
        if (grepl('PCOS|miscarriage|POP|endometriosis|Preeclampsia|leiomyoma_uterus', snakemake@input[[2]])) {
	if (grepl('PCOS', snakemake@input[[2]])) {s_pheno=  (1184 + 670 + 157 +658 +984 + 485 + 462 )/ (1184 + 670 + 157 +658 +984 + 485 + 462 + 5799 + 1379 +2807 +6774 +2963+ 407 + 96172)}
	if (grepl('miscarriage', snakemake@input[[2]])) {s_pheno=   49996 / ( 174109 + 49996)}
	if (grepl('POP', snakemake@input[[2]])) {s_pheno= 7053 / (57407 + 7053) }
	if (grepl('endometriosis', snakemake@input[[2]])) {s_pheno= 1496 / (192678 + 1496 )}
	if (grepl('Preeclampsia', snakemake@input[[2]])){ s_pheno= 4630/ (4630 + 373345)}
	if (grepl('leiomyoma_uterus', snakemake@input[[2]])){ s_pheno= ( 14569) / (85792 + 14569)}
	if (grepl('allPTD', snakemake@input[[1]])) {
	data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.067)
	} else if (grepl('postTerm', snakemake@input[[1]])) {
	data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.122) 
	} else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

	data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'cc', snp= temp_df$ID, s= s_pheno, MAF= temp_df$maf)

	} else {

        if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.067) 
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, MAF= temp_df$MAF, s= 0.122) 
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

	data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$N, type= 'quant', snp= temp_df$ID, MAF= temp_df$maf) 

	}
	myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
	if (length(myres)==1 ) { 
	PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, locus= locus)
        pph_list[[i]]= PPH
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, locus= locus)
	res_list[[i]]= res
	next 
	}
	PPH= data.frame(t(myres[[1]]))
	PPH$locus= locus        
	pph_list[[i]]= PPH
        res= myres[[2]]
        res$locus= locus
	res_list[[i]]= res
}
}

pph= data.frame(do.call('rbind', pph_list))
res= data.frame(do.call('rbind', res_list))


write.table(pph, snakemake@output[[1]], sep= '\t', row.names=F, col.names= T, quote=F)
write.table(res, snakemake@output[[2]], sep= '\t', row.names=F, col.names= T, quote=F)
R dplyr data.table Quant coloc From line 1 of colocalization/coloc.R 
14
15
script:
	'coloc.R'
SnakeMake From line 14 of colocalization/Snakefile 
23
24
25
26
27
28
29
30
31
run:
	df_list= list()
	for i in input:
		d= pd.read_csv(i, sep= '\t', header= 0)
		x= i.split('pph_')[1].replace('.txt', '')
		d['trait']= x
		df_list.append(d)
	d= pd.concat(df_list)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 23 of colocalization/Snakefile 
39
40
41
42
43
44
45
46
47
run:
        df_list= list()
        for i in input:
                d= pd.read_csv(i, sep= '\t', header= 0)
                x= i.split('results_')[1].replace('.txt', '')
                d['trait']= x
                df_list.append(d)
        d= pd.concat(df_list)
        d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 39 of colocalization/Snakefile 
61
62
script:
        'coloc_GA_vs_PTD.R'
SnakeMake From line 61 of colocalization/Snakefile 
75
76
script:
	'coloc_GA_vs_PTD_GW.R'
SnakeMake From line 75 of colocalization/Snakefile 
89
90
script:
        'coloc_BW_GA_GW.R'
SnakeMake From line 89 of colocalization/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
import pandas as pd
import numpy as np


def flip_beta(df):
        'Flip EFF and REF allele if REF> EFF. Flip beta direction with same condition. Assumed column names: beta, REF, EFF.'
        df['BETA']= np.where(df.REF>df.EFF, -1 * df.BETA, df.BETA)
        df['REF'], df['EFF']= np.where(df.REF> df.EFF, [df.EFF, df.REF], [df.REF, df.EFF])
        return df

def add_ID(x):
	x['REF']= np.where(x.REF.str.len() > x.EFF.str.len(), 'I', x.REF)
	x['EFF']= np.where(x.REF.str.len() < x.EFF.str.len(), 'I', x.EFF)
	x['REF']= np.where(x.EFF== 'I', 'D', x.REF)
	x['EFF']= np.where(x.REF== 'I', 'D', x.EFF)
	x['ID']= np.where(x.REF> x.EFF, x.CHR.apply(str) + ':' + x.POS.apply(str) + ':' + x.EFF + ':' + x.REF, x.CHR.apply(str) + ':' + x.POS.apply(str) + ':' + x.REF + ':' + x.EFF)
	x= flip_beta(x)
	return x


def format_df(x, reg):
	d= pd.read_csv(x, sep= ',', header= 0)
	d['chr']= d.chr.apply(str)
	d= pd.merge(d, reg, left_on= 'chr', right_on= 'CHR')
	d= d.loc[((d.pos >= d.pos1) & (d.pos<= d.pos2)), :]
	h1= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h1.coef', 'h1.se', 'h1.pval']]
	h1.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h2= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h2.coef', 'h2.se', 'h2.pval']]
	h2.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h3= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h3.coef', 'h3.se', 'h3.pval']]
	h3.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h1= add_ID(h1)
	h2= add_ID(h2)
	h3= add_ID(h3)
	h1.to_csv(snakemake.output[0], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h2.to_csv(snakemake.output[1], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h3.to_csv(snakemake.output[2], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	print('Completed file:' + x)

regions= pd.read_csv(snakemake.input[0], sep= '\t', header= 0)

format_df(snakemake.input[1], regions)
Python Pandas numpy From line 1 of effect_origin/format_CCHMC_haplotype.py 
29
30
script:
	'format_CCHMC_haplotype.py'
SnakeMake From line 29 of effect_origin/Snakefile 
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['contig', 'position', 'testedAllele', 'otherAllele', 'h.Bmnt', 'h.Bmnt.se', 'h.Bmnt.p', 'h.Bmt', 'h.Bmt.se', 'h.Bmt.p', 'h.Bft', 'h.Bft.se', 'h.Bft.p'])[['contig', 'position', 'testedAllele', 'otherAllele', 'h.Bmnt', 'h.Bmnt.se', 'h.Bmnt.p', 'h.Bmt', 'h.Bmt.se', 'h.Bmt.p', 'h.Bft', 'h.Bft.se', 'h.Bft.p']]
	d.columns= ['CHR', 'POS', 'EFF', 'REF', 'h2_beta', 'h2_se', 'h2_pvalue', 'h1_beta', 'h1_se', 'h1_pvalue', 'h3_beta', 'h3_se', 'h3_pvalue']
	h1= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'h1_beta', 'h1_se', 'h1_pvalue']]
	h1.columns= ['CHR', 'POS', 'EFF', 'REF', 'BETA', 'SE', 'pvalue']
	h2= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'h2_beta', 'h2_se', 'h2_pvalue']]
	h2.columns= ['CHR', 'POS', 'EFF', 'REF', 'BETA', 'SE', 'pvalue']
	h3= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'h3_beta', 'h3_se', 'h3_pvalue']]
	h3.columns= ['CHR', 'POS', 'EFF', 'REF', 'BETA', 'SE', 'pvalue']
	h1= add_ID(h1)
	h2= add_ID(h2)
	h3= add_ID(h3)
	h1.to_csv(output[0], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h2.to_csv(output[1], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h3.to_csv(output[2], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
SnakeMake From line 40 of effect_origin/Snakefile 
64
65
66
67
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d[['CHR', 'POS', 'REF', 'EFF']]= d.snp.str.split(':', expand= True)
	h1= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'beta_h1', 'se_h1', 'pvalue_h1']]
SnakeMake From line 64 of effect_origin/Snakefile 
90
91
shell:
	'/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 90 of effect_origin/Snakefile 
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
run:
	h1= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'Effect', 'StdErr', 'P-value', 'Direction', 'HetISq', 'HetPVal'])
	h2= pd.read_csv(input[1], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'Effect', 'StdErr', 'P-value', 'HetISq', 'HetPVal'])
	h3= pd.read_csv(input[2], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'Effect', 'StdErr', 'P-value', 'HetISq', 'HetPVal'])
	h1['N_cohorts']= 6 - h1['Direction'].apply(lambda x: str.count(x, '?'))
	h1.columns= ['MarkerName', 'Allele1', 'Allele2', 'beta_h1', 'se_h1', 'pvalue_h1', 'Direction_h1', 'HetISq_h1', 'HetPval_h1', 'n_cohorts']
	h1['beta_h1']= np.where(h1.Allele2> h1.Allele1, -1 * h1.beta_h1, h1.beta_h1)
	h2.columns= ['MarkerName', 'Allele1', 'Allele2', 'beta_h2', 'se_h2', 'pvalue_h2', 'HetISq_h2', 'HetPval_h2']
	h2['beta_h2']= np.where(h2.Allele2> h2.Allele1, -1 * h2.beta_h2, h2.beta_h2)
	h3.columns= ['MarkerName', 'Allele1', 'Allele2', 'beta_h3', 'se_h3', 'pvalue_h3', 'HetISq_h3', 'HetPval_h3']
	h3['beta_h3']= np.where(h3.Allele2> h3.Allele1, -1 * h3.beta_h3, h3.beta_h3)
	d= pd.merge(h1, h2[['MarkerName', 'beta_h2', 'se_h2', 'pvalue_h2', 'HetISq_h2', 'HetPval_h2']], on= ['MarkerName'], how= 'inner')
	d= pd.merge(d, h3[['MarkerName', 'beta_h3', 'se_h3', 'pvalue_h3', 'HetISq_h3', 'HetPval_h3']], on= ['MarkerName'], how= 'inner')
	d= d[['MarkerName', 'Allele1', 'Allele2', 'beta_h1', 'se_h1', 'pvalue_h1', 'Direction_h1', 'HetISq_h1', 'HetPval_h1', 'n_cohorts', 'beta_h2', 'se_h2', 'pvalue_h2', 'HetISq_h2', 'HetPval_h2', 'beta_h3', 'se_h3', 'pvalue_h3', 'HetISq_h3', 'HetPval_h3']]
	d['Allele1'], d['Allele2']= np.where(d.Allele2> d.Allele1, [d.Allele2, d.Allele1], [d.Allele1, d.Allele2])
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 101 of effect_origin/Snakefile 
 9
10
11
12
13
14
15
16
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'RSID', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue'])
	d.drop_duplicates(['ID'], keep= 'first', inplace= True)
	d.sort_values('pvalue', inplace= True, ascending= True)
	d= d.iloc[0:99999, :]
	d= d[['RSID', 'CHR', 'POS', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue']]
	d.columns= ['CHR', 'POS', 'RSID', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'PVALUE']
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 9 of EGG_sumstats/Snakefile 
25
26
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'RSID', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue'])
SnakeMake From line 25 of EGG_sumstats/Snakefile 
45
46
47
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'RSID', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue'])
        d.drop_duplicates(['ID'], keep= 'first', inplace= True)
SnakeMake From line 45 of EGG_sumstats/Snakefile 
65
66
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'RSID', 'REF', 'EFF', 'EAF', 'TOTALSAMPLESIZE', 'BETA', 'SE', 'pvalue'])
SnakeMake From line 65 of EGG_sumstats/Snakefile 
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	top_list= list()
	non_top_list= list()
	for lname in set(d.locus):
		df_temp= d.loc[d.locus== lname, :]
		df_temp.sort_values(['PP'], ascending= False, inplace= True)
		df_temp['PPcum']= df_temp.PP.cumsum()
		top_vars= df_temp.loc[df_temp.PPcum< 0.95, :]
		non_top= df_temp.loc[df_temp.PPcum>= 0.95, :]
		top_list.append(top_vars)
		non_top_list.append(non_top)
	top= pd.concat(top_list)
	non_top= pd.concat(non_top_list)
	top.to_csv(output[0], sep= '\t', header= True, index= False)
	non_top.to_csv(output[1], sep= '\t', header= True, index= False)
SnakeMake From line 30 of enrichment/Snakefile 
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
run:
	d= pd.read_csv(input[0], sep='\t', header= 0)
	pli= pd.read_csv(input[1], header= 0, sep= '\t', usecols= ['gene_id', 'gene', 'chromosome', 'start_position', 'end_position', 'pLI'])[['gene_id', 'gene', 'chromosome', 'start_position', 'end_position', 'pLI']]
	d= d.loc[d.nearestGene.isin(pli.gene.values), :]
	pli.columns= ['EID', 'gene', 'CHR', 'start', 'end', 'pLI']
	pli.dropna(subset= ['pLI'], inplace= True)
	pli_genes= pli.loc[pli.pLI>= 0.9, 'gene'].values.tolist()
	df= d.loc[d.nearestGene.isin(pli_genes), :]
	b= len(pli_genes) - df.shape[0]
	c= d.shape[0] - df.shape[0]
	d= df.shape[0]
	a= pli.shape[0] - b - d - c
	oddsratio, pvalue = st.fisher_exact([[a, b],[c, d]], alternative= 'greater')
	z= ['pli', a, b, c, d, (d / (d+c)), (b / (a + b)), oddsratio, pvalue]
	with open(output[0], 'w') as file_handler:
		file_handler.write('\t'.join([str(item) for item in z]) + '\n')
SnakeMake From line 54 of enrichment/Snakefile 
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	rna= pd.read_csv(input[1], sep= '\t', header= 0)
	rna['GA']= np.where(rna['Gene name'].isin(d.nearestGene.values), 1, 0)
	rna['NX_rk']= rna.groupby('Gene name')['NX'].rank('average', ascending= True)
	df_list= list()
	for tissue in set(rna.Tissue):
		ilist= rna.loc[((rna.GA== 1) & (rna.Tissue == tissue)), 'NX_rk']
		base= rna.loc[((rna.GA== 0) & (rna.Tissue == tissue)), 'NX_rk']
		mannw_pvalue= st.mannwhitneyu(ilist, base, alternative= 'greater')[1]
		i_median= np.median(ilist)
		base_median= np.median(base)
		df_list.append([tissue, i_median, base_median, mannw_pvalue])
	z= pd.DataFrame.from_records(df_list)
	z.to_csv(output[0], sep= '\t', header= ['tissue', 'i_listmedian', 'base_list_median', 'MannW_pvalue'], index= False)
SnakeMake From line 78 of enrichment/Snakefile 
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
run:
	pheno= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'nearestGene'])
	x= pd.read_csv(input[1], sep= '\t', header= None, names= ['CHR', 'start', 'end', 'gene', 'EnsembleID'])
	add= [line.strip() for line in open(input[2], 'r')]
	df= pheno.loc[pheno.nearestGene.isin(add), :]
	b= len(add) - df.shape[0]
	c= pheno.shape[0] - df.shape[0]
	d= df.shape[0]
	a= x.shape[0] - b - d - c
	oddsratio, pvalue = st.fisher_exact([[a, b],[c, d]], alternative= 'greater')
	z= ['dominant', a, b, c, d, (d / (d+c)), (b / (a + b)), oddsratio, pvalue]
	with open(output[0], 'w') as file_handler:
		file_handler.write('\t'.join([str(item) for item in z]) + '\n')
	rec= [line.strip() for line in open(input[3], 'r')]
	df= pheno.loc[pheno.nearestGene.isin(rec), :]
	b= len(rec) - df.shape[0]
	c= pheno.shape[0] - df.shape[0]
	d= df.shape[0]
	a= x.shape[0] - b - d - c
	oddsratio, pvalue = st.fisher_exact([[a, b],[c, d]], alternative= 'greater')
	z= ['recessive', a, b, c, d, (d / (d+c)), (b / (a + b)), oddsratio, pvalue]
	with open(output[0], 'a') as file_handler:
		file_handler.write('\t'.join([str(item) for item in z]) + '\n')
SnakeMake From line 103 of enrichment/Snakefile 
134
135
136
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'nearestGene'])
	stc= pd.read_csv(input[1], header= 0, sep= '\t', usecols= ['geneid', 'log2FoldChange', 'pvalue'])[['geneid', 'log2FoldChange', 'pvalue']]
SnakeMake From line 134 of enrichment/Snakefile 
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
run:
	pheno= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID', 'nearestGene'])
	x= pd.read_csv(input[1], sep= '\t', header= None, names= ['CHR', 'start', 'end', 'gene', 'EnsembleID'])
	lab= pd.read_csv(input[2], sep= '\t', header= 0)
	for i in set(lab.Cell_type):
		temp_df= lab.loc[lab.Cell_type== i, :]
		df= pheno.loc[pheno.nearestGene.isin(temp_df.gene_name.to_list()), :]
		b= len(temp_df.gene_name.to_list()) - df.shape[0]
		c= pheno.shape[0] - df.shape[0]
		d= df.shape[0]
		a= x.shape[0] - b - d - c
		oddsratio, pvalue = st.fisher_exact([[a, b],[c, d]], alternative= 'greater')
		z= [i, a, b, c, d, (d / (d+c)), (b / (a + b)), oddsratio, pvalue]
		with open(output[0], 'a') as file_handler:
			file_handler.write('\t'.join([str(item) for item in z]) + '\n')
	df= pheno.loc[pheno.nearestGene.isin(lab.gene_name.to_list()), :]
SnakeMake From line 159 of enrichment/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

df= fread(snakemake@input[[1]], select= c('RSID', 'BETA', 'SE', 'TOTALSAMPLESIZE', 'EAF'))

df= filter(df, !duplicated(RSID))

df$MAF= ifelse(df$EAF>0.5, 1 - df$EAF, df$EAF)

z= fread(snakemake@input[[2]])
z$n= 206
z$maf= ifelse(z$Freq< 0.5, 1 - z$Freq, z$Freq)
df= inner_join(df, z, by= c('RSID'= 'SNP'))

rm(z)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]


cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tprotein\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tprotein\n', file= snakemake@output[[2]])

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6


df= data.frame(df)

colocalization_eqtl= function(temp_df){
	protein= unique(temp_df$Gene)
        if (nrow(temp_df)== 0) {

        PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')

        } else {
	temp_df = filter(temp_df, SE>0, se> 0)

	if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$RSID, s= 0.067, MAF= temp_df$MAF)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$RSID, s= 0.122, MAF= temp_df$MAF)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N= temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$RSID, MAF= temp_df$MAF) }

        data2= list(beta= temp_df$b, varbeta= temp_df$se**2, N=temp_df$n, type= 'quant', snp= temp_df$RSID, MAF= temp_df$maf)
        myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
        if (length(myres)==1 ) { 
        PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
        } else {
        PPH= data.frame(t(myres[[1]]))
        PPH$protein= protein
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$protein= protein
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
}
}
}



mclapply(split(df, df$Gene), colocalization_eqtl, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of eQTLs/coloc_endometrium.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

df= fread(snakemake@input[[1]], select= c('ID', 'BETA', 'SE', 'TOTALSAMPLESIZE', 'EAF'))

df= filter(df, !duplicated(ID))

df$MAF= ifelse(df$EAF>0.5, 1 - df$EAF, df$EAF)

z= fread(snakemake@input[[2]], select= c('gene_id', 'ID', 'maf', 'slope', 'slope_se'))


z$n= with(z, ifelse(grepl('Ovary', snakemake@input[[2]]), 167, ifelse(grepl('Uterus', snakemake@input[[2]]), 269, 141)))


df= inner_join(df, z, by= 'ID')

rm(z)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]


cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tprotein\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tprotein\n', file= snakemake@output[[2]])

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6


df= data.frame(df)

colocalization_eqtl= function(temp_df){
	protein= unique(temp_df$gene_id)
        if (nrow(temp_df)== 0) {

        PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')

        } else {
	temp_df = filter(temp_df, SE>0, slope_se> 0)

	if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, s= 0.067, MAF= temp_df$MAF)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, s= 0.122, MAF= temp_df$MAF)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N= temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

        data2= list(beta= temp_df$slope, varbeta= temp_df$slope_se**2, N=temp_df$n, type= 'quant', snp= temp_df$ID, MAF= temp_df$maf)
        myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
        if (length(myres)==1 ) { 
        PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
        } else {
        PPH= data.frame(t(myres[[1]]))
        PPH$protein= protein
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$protein= protein
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
}
}
}



mclapply(split(df, df$gene_id), colocalization_eqtl, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of eQTLs/coloc_GTEx.R 
11
12
script:
        'coloc_endometrium.R'
SnakeMake From line 11 of eQTLs/Snakefile 
20
21
22
23
24
25
26
27
run:
	df_list= list()
	for i in input:
		d= pd.read_csv(i, header= 0, sep= '\t', usecols= ['ID'])
		df_list.append(d)
	x= pd.concat(df_list)
	x.drop_duplicates('ID', keep= 'first', inplace= True)
	x.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 20 of eQTLs/Snakefile 
36
37
38
39
40
41
42
43
44
45
46
run:
	x= ['ID', 'gene_id', 'maf', 'slope', 'slope_se']
	with open(output[0], mode="w") as file:
		file.write("\t".join(x) + "\n")
	d= pd.read_csv(input[1], sep= '\t', header= 0)
	for chunk in pd.read_csv(input[0], sep= '\t', header= 0, chunksize= 500000, compression= 'gzip', usecols= ['gene_id', 'variant_id', 'maf', 'slope', 'slope_se']) :
		chunk[['CHR', 'POS', 'REF', 'EFF', 'build']]= chunk.variant_id.str.split('_', expand= True)
		chunk['ID']= np.where(chunk.REF> chunk.EFF, chunk.CHR + ':' + chunk.POS + ':' + chunk.EFF + ':' + chunk.REF, chunk.CHR + ':' + chunk.POS + ':' + chunk.REF + ':' + chunk.EFF)
		chunk= chunk[['ID', 'gene_id', 'maf', 'slope', 'slope_se']]
		chunk= pd.merge(chunk, d, on= 'ID')
		chunk.to_csv(output[0], sep= '\t', header= False, index= False, mode= 'a')
SnakeMake From line 36 of eQTLs/Snakefile 
60
61
script:
        'coloc_GTEx.R'
SnakeMake From line 60 of eQTLs/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
import pandas as pd
import numpy as np


def flip_beta(df):
        'Flip EFF and REF allele if REF> EFF. Flip beta direction with same condition. Assumed column names: beta, REF, EFF.'
        df['BETA']= np.where(df.REF>df.EFF, -1 * df.BETA, df.BETA)
        df['REF'], df['EFF']= np.where(df.REF> df.EFF, [df.EFF, df.REF], [df.REF, df.EFF])
        return df

def add_ID(x):
        x['REF']= np.where(x.REF.str.len() > x.EFF.str.len(), 'I', x.REF)
        x['EFF']= np.where(x.REF.str.len() < x.EFF.str.len(), 'I', x.EFF)
        x['REF']= np.where(x.EFF== 'I', 'D', x.REF)
        x['EFF']= np.where(x.REF== 'I', 'D', x.EFF)
        x['ID']= np.where(x.REF> x.EFF, x.CHR.apply(str) + ':' + x.POS.apply(str) + ':' + x.EFF + ':' + x.REF, x.CHR.apply(str) + ':' + x.POS.apply(str) + ':' + x.REF + ':' + x.EFF)
        x= flip_beta(x)
        return x


def format_df(x):
	d= pd.read_csv(x, sep= ',', header= 0)
	d['chr']= d.chr.apply(str)
	d= d.loc[d.chr== '2', :]
	d= d.loc[d.pos== 113521754, :]
	h1= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h1.coef', 'h1.se', 'h1.pval']]
	h1.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h2= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h2.coef', 'h2.se', 'h2.pval']]
	h2.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h3= d.loc[:, ['chr', 'pos', 'ref', 'alt', 'h3.coef', 'h3.se', 'h3.pval']]
	h3.columns= ['CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']
	h1= add_ID(h1)
	h2= add_ID(h2)
	h3= add_ID(h3)
	h1.to_csv(snakemake.output[0], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h2.to_csv(snakemake.output[1], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	h3.to_csv(snakemake.output[2], sep= '\t', header= True, index= False, columns= ['ID', 'CHR', 'POS', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
	print('Completed file:' + x)


format_df(snakemake.input[0])
Python Pandas numpy From line 1 of fetal_SNP/format_CCHMC_haplotype.py 
 9
10
script:
        'format_CCHMC_haplotype.py'
SnakeMake From line 9 of fetal_SNP/Snakefile 
20
21
22
23
24
25
run:
	d= pd.read_csv(input[0], sep= ' ', header= 0)
	d['CHR']= 2
	d['POS']= 113521754
	d['REF']= 'C'
	d['EFF']= 'T'
SnakeMake From line 20 of fetal_SNP/Snakefile 
47
48
49
50
51
52
53
run:
        d= pd.read_csv(input[0], sep= ' ', header= 0)
        d[['CHR', 'POS', 'REF', 'EFF']]= d.snp.str.split(':', expand= True)
        h1= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'beta_h1', 'se_h1', 'pvalue_h1']]
        h1.columns= ['CHR', 'POS', 'EFF', 'REF', 'BETA', 'SE', 'pvalue']
        h2= d.loc[:, ['CHR', 'POS', 'EFF', 'REF', 'beta_h2', 'se_h2', 'pvalue_h2']]
        h2.columns= ['CHR', 'POS', 'EFF', 'REF', 'BETA', 'SE', 'pvalue']
SnakeMake From line 47 of fetal_SNP/Snakefile 
73
74
shell:
        '/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 73 of fetal_SNP/Snakefile 
84
85
86
87
88
run:
	h1= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'Effect', 'StdErr', 'P-value'])
	h1.columns= ['ID', 'EFF', 'REF', 'beta_MT', 'se_MT', 'pvalue_MT']
	h1['beta_MT']= np.where(h1.REF > h1.EFF, -1 * h1.beta_MT, h1.beta_MT)
	h2= pd.read_csv(input[1], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'Effect', 'StdErr', 'P-value'])
SnakeMake From line 84 of fetal_SNP/Snakefile 
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)




d= fread(snakemake@input[[1]])
d1= fread(snakemake@input[[2]])

d$beta= as.numeric(d$beta)
d$se= as.numeric(d$se)
d$pvalue= as.numeric(d$pvalue)

d1= filter(d1, PP.H4.abf> 0.75)
d= filter(d, pheno %in% d1$pheno_FINNGEN)

mani= fread(snakemake@input[[3]], select= c('phenocode', 'name'))
names(mani)= c('pheno', 'description')

d= inner_join(d, mani, by= 'pheno')

x= fread(snakemake@input[[4]])

x1= fread(snakemake@input[[5]])
x1= filter(x1, PP.H4.abf> 0.75)

x= filter(x, pheno %in% x1$pheno_PAN_UKBB)

mani=fread(snakemake@input[[6]], select= c('phenocode', 'trait_type', 'description'))
mani$pheno= paste(mani$trait_type, mani$phenocode, sep= '_')

x= inner_join(x, mani, by= 'pheno')

d$zscore= d$beta / d$se
x$zscore= x$beta / x$se

d= select(d, pheno, description, zscore, pvalue, trait)
x= select(x, pheno, description, zscore, pvalue, trait)
d= bind_rows(d, x)




d$zscore= ifelse(d$zscore> 10, 10, ifelse(d$zscore< -10, -10, d$zscore))

d$trait= ifelse(d$trait== 'Gestational duration', 'rs28654158 (gestational duration)', 'rs11708067 (birth weight)')

d$trait= factor(d$trait, levels= rev(c('rs28654158 (gestational duration)', 'rs11708067 (birth weight)')))

d$description= with(d, ifelse(grepl('Other diabetes', description), 'Other diabetes', description))

d$description= with(d, ifelse(grepl('Non-insulin-dep', description), 'Non-insulin dependent diabetes', description))
d$description= with(d, ifelse(grepl('Diabetes, varying def', description), 'Diabetes, wide', description))
d$description= with(d, ifelse(grepl('Intestinal adhesions', description), 'Intestinal adhesions', description))

d$description= with(d, ifelse(grepl('Type 2 diabetes, strict', description), 'Type 2 diabetes', description))

d$description= with(d, ifelse(grepl('Type 2 diabetes with other specified/multiple/unspecified complications', description), 'Type 2 diabetes with complications', description))

d$description= with(d, ifelse(grepl('Diabetes, insuline treatment', description), 'Diabetes, insuline treatment', description))

d$description= with(d, ifelse(grepl('Creatinine', description), 'Creatinine in urine', description))

ord <- hclust( dist(d$zscore, method = "euclidean"), method = "ward.D" )$order
d= d[ord, ]
d$description= factor(d$description, levels= unique(d$description))


p1= ggplot(d, aes(y= trait, x= description, fill= round(zscore), alpha= factor(as.numeric(pvalue< 5e-6)))) +
geom_tile(colour = "white", size= 1) +
theme_cowplot(font_size= 8) +
scale_alpha_discrete(guide=F, range= c(0.3, 1)) +
scale_fill_gradient2(low= colorBlindBlack8[3], high= colorBlindBlack8[8], mid= 'white', guide= F) +
theme(  axis.text.x= element_text(hjust= 1, angle= 45),
	axis.text.y= element_text(),
	axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        plot.margin= unit(c(t= 0, r= 0, b= 0, l= 0), unit= 'cm'),
	axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3)) +
coord_equal() +
labs(x = NULL, y = NULL)




ggsave(snakemake@output[[1]], p1, height= 100, width= 127, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')


p1= ggplot(d, aes(description, trait, fill= round(zscore), alpha= factor(as.numeric(pvalue< 5e-6)))) +
geom_tile(colour = "white", size= 1) +
theme_cowplot(font_size= 8) +
scale_alpha_discrete(guide=F) +
scale_fill_gradient2(low= colorBlindBlack8[3], high= colorBlindBlack8[8], mid= 'white') +
theme(axis.text.x = element_text(angle = 45, hjust = 1),
        axis.title.x = element_blank(),
        axis.title.y = element_blank(),
        plot.margin= margin(t= 0, r= 0, l= 0, b= 0, unit= 'pt')) +
scale_x_discrete(position = "top")


ggsave(snakemake@output[[3]], p1, height= 100, width= 140, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 2 of figures/ADCY5_effect_direction.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
library(dplyr)
library(data.table)
library(ggplot2)
library(cowplot)
library(ggrepel)
library(tidyr)
library(showtext)
colorBlindBlack8  <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')
as= 8
as1= 9
showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])
names(d)= c('CHR', 'POS', 'FST_EUR_AFR')

d1= fread(snakemake@input[[2]])
names(d1)= c('CHR', 'POS', 'FST_EUR_EAS')

d2= fread(snakemake@input[[3]])
names(d2)= c('CHR', 'POS', 'FST_AFR_EAS')

d= inner_join(d, d1, by= c('CHR', 'POS')) %>% inner_join(., d2, by= c('CHR', 'POS'))

rm(d1); rm(d2)
d$v_ids= paste(d$CHR, d$POS, sep= ':')

z= fread(snakemake@input[[4]])

zl= gather(z, control_set, v_ids, Set_1:Set_10000)

bw_pos= c(123065778)
ga_pos= c(123112292)

zl= inner_join(zl, d[, c('v_ids', 'FST_EUR_AFR', 'FST_EUR_EAS', 'FST_AFR_EAS')], by= 'v_ids')

zl= filter(zl, Input_SNP== '3:123065778' | Input_SNP== '3:123112292')

zl$haplotype= with(zl, ifelse(Input_SNP== '3:123065778', 'Birth weight', 'Gestational duration'))

zl= zl[!duplicated(zl$v_ids), ]

zl= data.frame(zl)
d= data.frame(d)
df_list= list()
r_num= 1
for (i in c('FST_EUR_AFR', 'FST_AFR_EAS', 'FST_EUR_EAS')){

ga_pvalue=wilcox.test(zl[zl$haplotype== 'Gestational duration', i], mu= d[d$v_ids== '3:123112292', i], alternative= 'less')$p.value

m1= d[d$v_ids== '3:123112292', i]
mc1= mean(zl[zl$haplotype== 'Gestational duration', i], na.rm=T)
medc1= median(zl[zl$haplotype== 'Gestational duration', i], na.rm=T)
prop_above= prop.table(table(d[d$v_ids== '3:123112292', i]> zl[zl$haplotype== 'Gestational duration', i]))[2]
temp_df= data.frame(haplotype= 'Gestational duration', ancestries= i, FST= m1, FST_mean_controls= mc1, FST_median_controls= medc1, pvalue= ga_pvalue)


ga_pvalue= wilcox.test(zl[zl$haplotype== 'Birth weight', i], mu= d[d$v_ids== '3:123065778', i], alternative= 'less')$p.value

medc1= median(zl[zl$haplotype== 'Birth weight', i], na.rm=T)
m1= d[d$v_ids== '3:123065778', i]
mc1= mean(zl[zl$haplotype== 'Birth weight', i], na.rm=T)

temp_df2= data.frame(haplotype= 'Birth weight', ancestries= i, FST= m1, FST_mean_controls= mc1, FST_median_controls= medc1, pvalue= ga_pvalue)
temp_df= rbind(temp_df, temp_df2)
df_list[[r_num]]= temp_df

r_num= r_num + 1
}

xp= do.call('rbind', df_list)

xp$enrichment= with(xp, FST / FST_median_controls)

bw= filter(zl, haplotype== 'Birth weight') %>% select(FST_EUR_AFR, FST_EUR_EAS, FST_AFR_EAS)
ga= filter(zl, haplotype== 'Gestational duration') %>% select(FST_EUR_AFR, FST_EUR_EAS, FST_AFR_EAS)

names(bw)= c('FST_EUR_AFR_bw', 'FST_EUR_EAS_bw', 'FST_AFR_EAS_bw')

df1= cbind(bw, ga)

ga_fst= d[d$v_ids== '3:123112292', 'FST_EUR_AFR']
bw_fst= d[d$v_ids== '3:123065778', 'FST_EUR_AFR']
ga_fst_pvalue= xp[xp$haplotype== 'Gestational duration' & xp$ancestries== 'FST_EUR_AFR', 'enrichment']
bw_fst_pvalue= xp[xp$haplotype== 'Birth weight' & xp$ancestries== 'FST_EUR_AFR', 'enrichment']


p1= ggplot(df1, aes(x=x) ) +
  geom_density( aes(x = FST_EUR_AFR, y = ..density..), fill= colorBlindBlack8[4], colour= colorBlindBlack8[4]) +
annotate('text', x=0.6, y= 10, label="Gestational \nduration", color= colorBlindBlack8[4], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=0.6, y= -10 - 0.5, label="Birth weight", color= colorBlindBlack8[2], size= as1/ .pt, fontface = 'bold') +
  annotate('text', x=ga_fst, y=5 + 0.5, label="rs28654158", color= colorBlindBlack8[4], size= as/ .pt) +
  annotate('text', x=bw_fst, y= -10 - 0.5, label="rs11708067", color= colorBlindBlack8[2], hjust= 0, size= as/ .pt) +
  annotate('text', x= 0.6, y= 1, label= paste0('Enrichment x', round(ga_fst_pvalue, 1)), color= colorBlindBlack8[4], size= as/ .pt) +
  annotate('text', x= 0.6, y= -1, label= paste0('Enrichment x', round(bw_fst_pvalue, 1)), color= colorBlindBlack8[2], size= as/ .pt) +
  geom_density(aes(x = FST_EUR_AFR_bw, y = -..density..), fill= colorBlindBlack8[2], colour= colorBlindBlack8[2]) +
  theme_cowplot(font_size = 8) +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(limits= c(-11, 11), breaks= c(-10, -5, 0, 5, 10), labels= c(10, 5, 0, 5, 10)) +
  xlab("Fst Africans - Europeans") +
ylab('Density') +
geom_segment(aes(x = ga_fst, y = 0, xend = ga_fst, yend = 5)) +
geom_segment(aes(x = bw_fst, y = 0, xend = bw_fst, yend = -10))+
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))



ggsave(snakemake@output[[1]], p1, width= 63, height= 63, units= 'mm', dpi= 300)

ga_fst= d[d$v_ids== '3:123112292', 'FST_EUR_EAS']
bw_fst= d[d$v_ids== '3:123065778', 'FST_EUR_EAS']
ga_fst_pvalue= xp[xp$haplotype== 'Gestational duration' & xp$ancestries== 'FST_EUR_EAS', 'enrichment']
bw_fst_pvalue= xp[xp$haplotype== 'Birth weight' & xp$ancestries== 'FST_EUR_EAS', 'enrichment']

p1= ggplot(df1, aes(x=x) ) +
  geom_density( aes(x = FST_EUR_EAS, y = ..density..), fill= colorBlindBlack8[4], colour= colorBlindBlack8[4]) +
annotate('text', x=0.57, y= 9, label="Gestational \nduration", color= colorBlindBlack8[4], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=0.57, y= -10, label="Birth weight", color= colorBlindBlack8[2], size= as1/ .pt, fontface = 'bold') +
  annotate('text', x=ga_fst, y= 10 + 0.5, label="rs28654158", color= colorBlindBlack8[4], hjust= 0, size= as/ .pt) +
  annotate('text', x=bw_fst, y= -5 - 0.5, label="rs11708067", color= colorBlindBlack8[2], size= as/ .pt) +
  annotate('text', x= 0.6, y= 1, label= paste0('Enrichment x', round(ga_fst_pvalue, 1)), color= colorBlindBlack8[4], size= as/ .pt) +
  annotate('text', x= 0.6, y= -1, label= paste0('Enrichment x', round(bw_fst_pvalue, 1)), color= colorBlindBlack8[2], size= as/ .pt) +
  geom_density( aes(x = FST_EUR_EAS_bw, y = -..density..), fill= colorBlindBlack8[2], colour= colorBlindBlack8[2]) +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(limits= c(-11, 11), breaks= c(-10, -5, 0, 5, 10), labels= c(10, 5, 0, 5, 10)) +
  theme_cowplot(font_size = 8) +
  xlab("Fst East Asians - Europeans") +
ylab('Density') +
geom_segment(aes(x = ga_fst, y = 0, xend = ga_fst, yend = 10)) +
geom_segment(aes(x = bw_fst, y = 0, xend = bw_fst, yend = -5)) +
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[2]], p1, width= 63, height= 63, units= 'mm', dpi= 300)

ga_fst= d[d$v_id== '3:123112292', 'FST_AFR_EAS']
bw_fst= d[d$v_id== '3:123065778', 'FST_AFR_EAS']
ga_fst_pvalue= xp[xp$haplotype== 'Gestational duration' & xp$ancestries== 'FST_AFR_EAS', 'enrichment']
bw_fst_pvalue= xp[xp$haplotype== 'Birth weight' & xp$ancestries== 'FST_AFR_EAS', 'enrichment']

p1= ggplot(df1, aes(x=x) ) +
geom_density( aes(x = FST_AFR_EAS, y = ..density..), fill= colorBlindBlack8[4], colour= colorBlindBlack8[4]) +
annotate('text', x=0.72, y=7, label="Gestational \nduration", color= colorBlindBlack8[4], size= as1/ .pt, fontface = 'bold') + 
annotate('text', x=0.72, y= -7, label="Birth weight", color= colorBlindBlack8[2], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=ga_fst, y=5 + 0.5, label="rs28654158", color= colorBlindBlack8[4], size= as/ .pt) +
annotate('text', x=bw_fst, y= -5 - 0.5, label="rs11708067", color= colorBlindBlack8[2], hjust= 0, size= as/ .pt) +
annotate('text', x= 0.75, y= 1, label= paste0('Enrichment x', round(ga_fst_pvalue, 1)), color= colorBlindBlack8[4], size= as/ .pt) +
annotate('text', x= 0.75, y= -1, label= paste0('Enrichment x', round(bw_fst_pvalue, 1)), color= colorBlindBlack8[2], size= as/ .pt) +
geom_density( aes(x = FST_AFR_EAS_bw, y = -..density..), fill= colorBlindBlack8[2], colour= colorBlindBlack8[2]) +
scale_x_continuous(expand= c(0, 0)) +
theme_cowplot(font_size = 8) +
xlab("Fst Africans - East Asians") +
scale_y_continuous(limits= c(-11, 11), breaks= c(-10, -5, 0, 5, 10), labels= c(10, 5, 0, 5, 10)) +
ylab('Density') +
geom_segment(aes(x = ga_fst, y = 0, xend = ga_fst, yend = 5)) +
geom_segment(aes(x = bw_fst, y = 0, xend = bw_fst, yend = -5))+
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))



ggsave(snakemake@output[[3]], p1, width= 63, height= 63, units= 'mm', dpi= 300)

fwrite(df1, snakemake@output[[4]], sep= '\t')
fwrite(xp, snakemake@output[[5]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel showtext From line 1 of figures/ADCY5_FST_AFR_EUR.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
library(dplyr)
library(data.table)
library(ggplot2)
library(cowplot)
library(ggrepel)
library(tidyr)
library(showtext)
colorBlindBlack8  <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])
names(d)[8]= 'phenocode'
mani= fread(snakemake@input[[2]])

trait_list= c('biomarkers', 'continuous', 'icd10')
mani= mani[mani$trait_type %in% trait_list, ]

mani= filter(mani, saige_heritability_EUR> 0.01)
mani= mani[order(mani$saige_heritability_EUR, decreasing= TRUE), ]
mani= mani[!duplicated(mani$phenocode), ]

mani$phenocode= paste(mani$trait_type, mani$phenocode, sep= '_')
mani= mani[, c('phenocode', 'description')]
mani= mani[!duplicated(mani$description), ]

d= inner_join(d, mani[, c('description', 'phenocode')], by= 'phenocode')
d$cohort= 'UKBB'

x= fread(snakemake@input[[3]])
names(x)[8]= 'phenocode'
mani= fread(snakemake@input[[4]])
mani= mani[, c('phenocode', 'name')]
names(mani)= c('phenocode', 'description')
mani= mani[!duplicated(mani$description), ]

x= inner_join(x, mani, by= 'phenocode')
x$cohort= 'FINNGEN'

d= rbind(d, x)
d= d[order(d$PP.H4.abf, decreasing= F), ]
d= filter(d, PP.H4.abf> 0.01, PP.H4.abf + PP.H3.abf> 0.75)

d$preg_trait= factor(d$preg_trait)
empty_bar <- 5
to_add <- data.frame( matrix(NA, empty_bar*nlevels(d$preg_trait), ncol(d)) )
colnames(to_add) <- colnames(d)
to_add$preg_trait <- rep(levels(d$preg_trait), each=empty_bar)
d <- rbind(d, to_add)
d <- d %>% arrange(preg_trait)


d$id= seq(1, nrow(d))

label_data= d
number_of_bar <- nrow(label_data)
angle <-  90 - 360 * (label_data$id-0.5) /number_of_bar
label_data$hjust<-ifelse( angle < -90, 1, 0)


label_data$angle<-ifelse(angle < -90, angle+180, angle)

#d$id= factor(d$id, levels= d$id[order(d$PP.H4.abf)])

base_data= d %>%
  group_by(preg_trait) %>%
  filter(is.na(PP.H4.abf)) %>%
  summarize(start=min(id), end=max(id) ) %>%
  rowwise() %>%
  mutate(title=mean(c(start, end)))

arc100= rep(1, 2)
arc75= rep(0.75, 2)
arc50= rep(0.50, 2)
arc25= rep(0.25, 2)

label_data$description= with(label_data, ifelse(grepl('Other diabetes', description), 'Other diabetes', description))

label_data$description= with(label_data, ifelse(grepl('Non-insulin-dep', description), 'Non-insulin dependent diabetes', description))
label_data$description= with(label_data, ifelse(grepl('Diabetes, varying def', description), 'Diabetes, wide', description))
label_data$description= with(label_data, ifelse(grepl('Intestinal adhesions', description), 'Intestinal adhesions', description))

label_data$description= with(label_data, ifelse(grepl('Type 2 diabetes, strict', description), 'Type 2 diabetes', description))

label_data$description= with(label_data, ifelse(grepl('Type 2 diabetes with other specified/multiple/unspecified complications', description), 'Type 2 diabetes with complications', description))

label_data$description= with(label_data, ifelse(grepl('and lymph nodes, not elsewhere classified', description), 'Diseases of veins', description))

label_data$description= with(label_data, ifelse(grepl('Diabetes, insuline treatment', description), 'Diabetes, insuline treatment', description))

label_data$description= with(label_data, ifelse(grepl('Creatinine', description), 'Creatinine in urine', description))

p1= ggplot(d, aes(as.factor(id), PP.H4.abf, fill= preg_trait, alpha= PP.H4.abf)) +
geom_bar(stat="identity", colour= NA) +
scale_alpha_continuous(range= c(0.4, 1), guide= F) +
geom_segment(data=base_data, aes(x = end, y = arc100, xend = start, yend = arc100), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc75, xend = start, yend = arc75), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc50, xend = start, yend = arc50), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc25, xend = start, yend = arc25), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  annotate("text", x = ((base_data$end[1] + base_data$start[1]) / 2), y = c((0.25 + 0.075) , (0.50 + 0.075), (0.75 + 0.075) , (1 + 0.075)), label = c("0.25", "0.50", "0.75", "1") , color="grey", size=2.5 , angle=13, fontface="bold", hjust= 0.5) +
   annotate("text", x = ((base_data$end[2] + base_data$start[2]) / 2), y = c((0.25 + 0.075) , (0.50 + 0.075), (0.75 + 0.075) , (1 + 0.075) ), label = c("0.25", "0.50", "0.75", "1") , color="grey", size=2.5, angle=13, fontface="bold", hjust=0.5) +
ylim(-0.2, 2) + # Limits of the plot = very important. The negative value controls the size of the inner circle, the positive one is useful to add size over each bar
theme_cowplot() +
scale_fill_manual(values=colorBlindBlack8[c(2,4)], guide= F) +
scale_colour_manual(values=colorBlindBlack8[c(2,4)], guide= F) +
#    plot.margin = margin(t= -200, r= -40, b= -200, l=-70, unit= 'mm')   ) +
labs(x=NULL, y=NULL)  +
  coord_polar(start = 0) +
geom_text(data=filter(label_data, PP.H4.abf> 0.75), aes(x= factor(id), y=PP.H4.abf + 0.01, label=description, hjust=hjust), color="black", fontface="bold",alpha=0.6, size=6/ .pt, angle= filter(label_data, PP.H4.abf> 0.750)$angle, inherit.aes = FALSE) +
theme(axis.line=element_blank(),axis.text.x=element_blank(),
          axis.text.y=element_blank(),axis.ticks=element_blank(),
          axis.title.x=element_blank(),
          axis.title.y=element_blank(),legend.position="none",
          panel.background=element_blank(),panel.border=element_blank(),panel.grid.major=element_blank(),
          panel.grid.minor=element_blank(),plot.background=element_blank(),
axis.ticks.length = unit(0, "mm"))

ggsave(snakemake@output[[1]], plot= p1, width= 127, height= 127, dpi= 300, units= 'mm')

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel showtext From line 1 of figures/ADCY5_pheWAS.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
library(scales)
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(tidyverse)
library(fmsb)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

flist= snakemake@input #list.files('/mnt/hdd/common/pol/metaGWAS/colocalization/GAraw/', 'pph_BW_', full.names=T)

funk= function(x){
d= fread(x)
d= filter(d, PP.H1.abf + PP.H2.abf + PP.H3.abf + PP.H4.abf + PP.H0.abf> 0)
fname= gsub('.txt', '', gsub('pph_', '', unlist(strsplit(x, '/'))[9]))
d= separate(d, locus, into= c('chrom', 'locus'), sep= '_')
d$sloc= d$PP.H4.abf + d$PP.H3.abf
d= select(d, PP.H4.abf, sloc, locus)

names(d)= c(fname, paste0(fname, '_sloc'), 'locus')
return(d)
}

d= lapply(flist, funk)

d= reduce(d, full_join, by = "locus")

d= arrange(d, BW_maternal_effect)

# Spider plot maternal

x= as.data.frame(matrix(d$BW_maternal_effect, ncol= nrow(d)))
x=rbind(x, as.data.frame(matrix(d$BW_maternal_effect_sloc, ncol= nrow(d))))
names(x)= d$locus

rownames(x)= c('BW maternal effect', 'BW maternal effect ')

x= rbind(rep(1,nrow(d)) , rep(0,nrow(d)) , x)


png(snakemake@output[[1]], width= 60, height= 60, res= 300, units= 'mm')
par(mar=c(0,0,0,0))

radarchart(x, axistype= 0, 

    #custom polygon
    pcol= c(colorBlindBlack8[4], colorBlindBlack8[2]) , pfcol= c(alpha(colorBlindBlack8[4], 0.4), alpha(colorBlindBlack8[2], 0.4)) , plwd=1, pty= 32, plty= 1,
    #custom the grid
    cglcol="grey", cglty=1, axislabcol="#525252", caxislabels= seq(0, 1, 0.25), caxisoffset= 0.1, cglwd=0.8, calcex= 0.4,

    #custom labels
    vlcex= 0.43
    )

dev.off()


# Spider plot fetal

x= as.data.frame(matrix(d$BW_fetal_effect, ncol= nrow(d)))
x=rbind(x, as.data.frame(matrix(d$BW_fetal_effect_sloc, ncol= nrow(d))))
names(x)= d$locus

rownames(x)= c('BW fetal effect', 'BW fetal effect ')

x= rbind(rep(1,nrow(d)) , rep(0,nrow(d)) , x)


png(snakemake@output[[2]], width= 60, height= 60, res= 300, units= 'mm')
par(mar=c(0,0,0,0))

radarchart(x, axistype= 0,

    #custom polygon
    pcol= c(colorBlindBlack8[4], colorBlindBlack8[2]) , pfcol= c(alpha(colorBlindBlack8[4], 0.4), alpha(colorBlindBlack8[2], 0.4)) , plwd=1, pty= 32, plty= 1,
    #custom the grid
    cglcol="grey", cglty=1, axislabcol="#525252", caxislabels= seq(0, 1, 0.25), caxisoffset= 0.1, cglwd=0.8, calcex= 0.4,

    #custom labels
    vlcex= 0.43
    )

dev.off()
R tidyverse dplyr data.table tidyr cowplot ggrepel knitr scales showtext From line 1 of figures/BW_coloc_spider.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

as= 8
as1= 8

z= fread(snakemake@input[[1]])
z$chr= as.numeric(gsub('chr', '', z$chr))
z$chr= as.character(z$chr)
z$locus= 1:nrow(z)


funk= function(infile){
d= fread(infile)
names(d)[1:11]= names(d)[2:12]
d=d[, 1:11]

d= filter(d, p<5e-6)

d$bC= ifelse(d$b< 0, -1 * d$bC, d$bC)
d$b= ifelse(d$b< 0, -1 * d$b, d$b)

d= separate(d, SNP, into= c('chr', 'POS', 'REF', 'EFF'), sep= ':')

d$POS= as.numeric(d$POS)
d$chr= as.character(d$chr)
d$GWAS= ifelse(grepl('BW_maternal_effect_GA', infile), 'BW_maternal_GA', 'BW_fetal_GA')
d= inner_join(d, z, on= 'chr') 
d= d %>% filter(POS>= start, POS< stop)

d= group_by(d, locus) %>% arrange(p) %>% filter(row_number()== 1)

return(d)

}

df_list= lapply(snakemake@input[grepl('BW', snakemake@input)], funk)

d= do.call('rbind', df_list)

d$beta_dif= with(d, (bC - b) / b)


mor= filter(d, GWAS== 'BW_maternal_GA') %>% pull(beta_dif)
barn= filter(d, GWAS== 'BW_fetal_GA') %>% pull(beta_dif)

p1= ggplot() +
geom_density( mapping=aes(x = mor, y = ..density..), fill= colorBlindBlack8[3], colour= colorBlindBlack8[3]) +
annotate('text', x=0.1, y= 3, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=0.1, y= -15, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
geom_density(mapping= aes(x = barn, y = -..density..), fill= colorBlindBlack8[8], colour= colorBlindBlack8[8]) +
  theme_cowplot(font_size = 8) +
scale_x_continuous(expand= c(0, 0)) +
  xlab("Relative difference in effect size on \nbirth weight after conditioning") +
ylab('Density') +
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

p1= ggplot(d, aes(beta_dif, group= GWAS, fill= GWAS)) +
geom_hline(yintercept= 0, colour= 'black') +
geom_density(color= NA) +
annotate('text', x=-0.75, y= 1, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=0.1, y= 15, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
theme_cowplot(font_size= 8) +
#scale_colour_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(expand=c(0, 0.5)) +
  xlab("Relative difference in effect size on \nbirth weight after conditioning") +
ylab('Density') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))

ggsave(snakemake@output[[3]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/BW_conditioning.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

as= 8
as1= 8

z= fread(snakemake@input[[1]])

z$SNP= with(z, ifelse(ref> eff, paste(chr, pos, eff, ref, sep= ':'), paste(chr, pos, ref, eff, sep= ':')))

funk= function(infile){
d= fread(infile)
names(d)[1:11]= names(d)[2:12]
d=d[, 1:11]

d$bC= ifelse(d$b< 0, -1 * d$bC, d$bC)
d$b= ifelse(d$b< 0, -1 * d$b, d$b)

d$GWAS= ifelse(grepl('BW_maternal_effect_GA', infile), 'BW_maternal_GA', 'BW_fetal_GA')

var= ifelse(grepl('BW_maternal_effect_GA', infile), 'Maternal Only', 'Fetal Only')
temp_z= z[z$origin== var, ]

d= filter(d, SNP %in% temp_z$SNP)

return(d)

}

df_list= lapply(snakemake@input[grepl('BW', snakemake@input)], funk)

d= do.call('rbind', df_list)

d$beta_dif= with(d, (bC - b) / b)


mor= filter(d, GWAS== 'BW_maternal_GA') %>% pull(beta_dif)
barn= filter(d, GWAS== 'BW_fetal_GA') %>% pull(beta_dif)

p1= ggplot() +
geom_density( mapping=aes(x = mor, y = ..density..), fill= colorBlindBlack8[3], colour= colorBlindBlack8[3]) +
annotate('text', x= 0.1, y= 3, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x= 0.1, y= -15, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
geom_density(mapping= aes(x = barn, y = -..density..), fill= colorBlindBlack8[8], colour= colorBlindBlack8[8]) +
  theme_cowplot(font_size = 8) +
scale_x_continuous(expand= c(0, 0)) +
  xlab("Relative difference in effect size on \nbirth weight after conditioning") +
ylab('Density') +
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

p1= ggplot(d, aes(beta_dif, group= GWAS, fill= GWAS)) +
geom_hline(yintercept= 0, colour= 'black') +
geom_density(color= NA) +
annotate('text', x=-0.55, y= 1, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=0.1, y= 10, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
theme_cowplot(font_size= 8) +
#scale_colour_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(expand=c(0, 0.5)) +
  xlab("Relative difference in effect size on \nbirth weight after conditioning") +
ylab('Density') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))

ggsave(snakemake@output[[3]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/BW_conditioning_top.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')


showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)



d= fread(snakemake@input[[1]])
d= filter(d, grepl('GAraw', p1), grepl('BW', p2))
d$p1= 'Gestational duration (maternal)'


x= fread(snakemake@input[[2]])

x= filter(x, grepl('GA_fetal', p1), grepl('BW', p2))
x$p1= 'Gestational duration (fetal)'



d= rbind(d, x)

d$p2= gsub('.txt.sumstats.gz', '', apply(d[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))
d$trait= d$p2
d$trait= with(d, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal \nonly',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal \nonly',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis'))))))))))))))))))))))



p1= ggplot(d, aes(trait, rg, colour= p1)) +
geom_pointrange(aes(ymin= rg - se * 1.96, ymax= rg + se * 1.96), position = position_dodge(0.3), width = 1/10, size= 0.4, fatten= 0.6) +
scale_colour_manual(values= colorBlindBlack8[c(8,3)], guide= FALSE) +
theme_cowplot(font_size= 8) +
scale_y_continuous(limits= c(-0.2, 0.8), breaks= seq(-0.2, 0.8, 0.2)) +
ylab('Genetic correlation') +
xlab('Effect on birth weight') +
geom_hline(yintercept= 0, size= 0.3) +
geom_hline(yintercept= c(-0.2, seq(0.2, 0.8, 0.2)), colour= 'grey', linetype= 'dashed', alpha= 0.5, size= 0.2) +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
	axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[1]], plot= p1, width= 60, height= 60, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/BW_genetic_correlations.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
library(data.table)
library(dplyr)
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])
x= fread(snakemake@input[[2]])
d= inner_join(d, x, by= 'Name')


d= d[sample(nrow(d)),]

d= d[order(d$Category, decreasing= F), ]

d$Name= factor(d$Name, levels= unique(d$Name))

d$Name2= gsub('_', ' ', gsub("^.*\\.","", d$Name))
d$Name2= factor(d$Name2, levels= unique(d$Name2))

p1= ggplot(d, aes(Name2, -log10(Coefficient_P_value), colour= Category, fill= Category)) + 
geom_point(size= 2, shape= 21, stroke= 0.1) +
xlab('Tissues') +
ylab('-log10(Enrichment)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= -log10(0.05), colour= '#d9d9d9') +
theme(axis.text.x = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05),
	panel.grid.major.x= element_blank(),
	legend.position="none") +
geom_text_repel(data= filter(d, Coefficient_P_value< 0.05), aes(Name2, -log10(Coefficient_P_value), colour= Category, label= Name2, show_guide = FALSE))


ggsave(snakemake@output[[1]], plot= p1, width= 120, height= 90, units= 'mm', dpi= 300)

p2= ggplot(d, aes(Name2, -log10(Coefficient_P_value), colour= Category, fill= Category)) +
geom_point(size= 2, shape= 21, stroke= 0.1) +
xlab('Tissues') +
ylab('-log10(Enrichment)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= -log10(0.05), colour= '#d9d9d9') +
theme(axis.text.x = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05),
        panel.grid.major.x= element_blank()) +
geom_text_repel(data= filter(d, Coefficient_P_value< 0.05), aes(Name2, -log10(Coefficient_P_value), colour= Category, label= Name2), show_guide = FALSE)

ggsave(snakemake@output[[2]], plot= p2, width= 120, height= 90, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table cowplot ggrepel showtext From line 1 of figures/cell_type_enrichment.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(ggtern)
options(warn=-1)



colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")


font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)



shbg= fread(snakemake@input[[1]])

testo= fread(snakemake@input[[2]])


#shbg$locus= gsub("^.*\\_","", shbg$locus)
#testo$locus= gsub("^.*\\_","", testo$locus)




colT= colorBlindBlack8[4]
colR= colorBlindBlack8[1]
colL= colorBlindBlack8[2]

shbg$One_or_Other= shbg$PP.H0.abf + shbg$PP.H1.abf + shbg$PP.H2.abf
shbg$coloc= shbg$PP.H4.abf
shbg$shared_locus= shbg$PP.H3.abf

p1= ggtern(shbg, aes(One_or_Other, coloc, shared_locus)) +
geom_point(colour= colorBlindBlack8[8], fill= colorBlindBlack8[8], shape= 21) +
scale_alpha_continuous(range= c(0.6, 1), guide= F) +
scale_size_continuous(range= c(.001, 10), guide= F) +
theme_custom(tern.plot.background = NULL, tern.panel.background = 'white', col.T = colT, col.L = colL, col.R = colR, col.grid.minor = "white") +
Tarrowlab("Probability of shared causal variant") +
Larrowlab("Probability of locus not shared") +
Rarrowlab("Probability of shared locus (distinct causal variant)")  +
theme_showarrows()  +
theme_notitles() +
theme(text=element_text(family="arial", size= 10),
        tern.axis.arrow.T = element_blank(),
        tern.axis.arrow.L = element_blank(),
        tern.axis.arrow.R = element_blank(),
        tern.axis.text.T = element_text(color = colT),
        tern.axis.text.L = element_text(color = colL),
        tern.axis.text.R = element_text(color = colR),
        tern.axis.arrow.text.T = element_text(color = colT),
        plot.margin = margin(0, 0, 0, 0, "cm"),
        tern.axis.arrow.text.L = element_text(color = colL),
        tern.axis.arrow.text.R = element_text(color = colR),
        tern.panel.grid.major = element_line(linetype = 6, size = 0.3))

testo$One_or_Other= testo$PP.H0.abf + testo$PP.H1.abf + testo$PP.H2.abf
testo$coloc= testo$PP.H4.abf
testo$shared_locus= testo$PP.H3.abf

p2= ggtern(testo, aes(One_or_Other, coloc, shared_locus)) +
geom_point(colour= colorBlindBlack8[8], fill= colorBlindBlack8[8], shape= 21) +
scale_alpha_continuous(range= c(0.6, 1), guide= F) +
scale_size_continuous(range= c(.001, 10), guide= F) +
theme_custom(tern.plot.background = NULL, tern.panel.background = 'white', col.T = colT, col.L = colL, col.R = colR, col.grid.minor = "white") +
Tarrowlab("Probability of shared causal variant") +
Larrowlab("Probability of locus not shared") +
Rarrowlab("Probability of shared locus (distinct causal variant)")  +
theme_showarrows()  +
theme_notitles() +
theme(text=element_text(family="arial", size= 10),
        tern.axis.arrow.T = element_blank(),
        tern.axis.arrow.L = element_blank(),
        tern.axis.arrow.R = element_blank(),
        tern.axis.text.T = element_text(color = colT),
        tern.axis.text.L = element_text(color = colL),
        tern.axis.text.R = element_text(color = colR),
        tern.axis.arrow.text.T = element_text(color = colT),
        plot.margin = margin(0, 0, 0, 0, "cm"),
        tern.axis.arrow.text.L = element_text(color = colL),
        tern.axis.arrow.text.R = element_text(color = colR),
        tern.panel.grid.major = element_line(linetype = 6, size = 0.3))

ggsave(snakemake@output[[1]], plot= p1, width= 95, height= 95, units= 'mm', dpi= 300)

ggsave(snakemake@output[[2]], plot= p2, width= 95, height= 95, units= 'mm', dpi= 300)
R dplyr data.table tidyr cowplot ggrepel knitr showtext ggtern From line 1 of figures/coloc_sex_hormones.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(ggdendro)
library(gridExtra)
library(dendextend)
library(plyr)
library(ggtree)
library(scales)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[1]])


x= fread(snakemake@input[[2]], select= c('nearestGene', 'RSID'))

d= inner_join(d, x, by= c('rsid'= 'RSID'))

d$GENE= d$nearestGene
d$GENE= with(d, ifelse(GENE== 'CDC42', 'CDC42/ WNT4', ifelse(GENE== 'HIVEP3', 'HIVEP3/ EDN2', ifelse(GENE== 'TET3', 'TET3/ DGUOK-AS1', ifelse(GENE== 'TCEA2', 'TCEA2/ OPRL1', GENE)))))
d$nearestGene= d$GENE

d$nearestGene= with(d, ifelse(rsid== 'rs3129768', 'HLA-DQA1', ifelse(rsid== 'rs5991030', 'AGTR2', ifelse(rsid== 'rs5930554', 'RAP2C', nearestGene)))) 

d$nearestGene= with(d, ifelse(rsid== 'rs6780427', 'KCNAB1', nearestGene))
d$nearestGene= with(d, ifelse(rsid== 'rs6879092', 'EBF1', nearestGene))


d$nearestGene= gsub(' ', '', d$nearestGene)
d$nearestGene= paste0("(", d$nearestGene, ")")
d$rsid_lab= with(d, paste(rsid, nearestGene))

d$beta_PT= with(d, ifelse(beta_MT<0, -1 * beta_PT, beta_PT))
d$beta_MNT= with(d, ifelse(beta_MT<0, -1 * beta_MNT, beta_MNT))
d$beta_MT= with(d, ifelse(beta_MT<0, -1 * beta_MT, beta_MT))

d= gather(d, haplotype, beta, c('beta_MT', 'beta_MNT', 'beta_PT'))

max_beta= max(d$beta)
min_beta= min(d$beta)


d$haplotype= with(d, ifelse(haplotype== 'beta_MT', 'Maternal\ntransmitted', ifelse(haplotype== 'beta_MNT', 'Maternal\nnon-transmitted', 'Paternal\ntransmitted')))
d$rsid_lab= factor(d$rsid_lab, levels= unique(d$rsid_lab))


d$class_name= factor(d$class_name, levels= c("Maternal", "MF SD", "MF OD", "Fetal MatT", "Fetal"))

d= d %>% arrange(class_name, desc(probability)) %>% ungroup()
d$rsid_lab= factor(d$rsid_lab, levels= unique(d$rsid_lab))

labs <- sapply(
  strsplit(levels(d$rsid_lab), " "), 
  function(x) parse(text = paste0(x[1], "~italic('", x[2], "')"))
)

p1= ggplot(d, aes(rsid_lab, haplotype, fill= beta)) +
  theme_cowplot(8) +
  geom_tile() +
  #scale_fill_gradient2(low= colorBlindBlack8[4], high= colorBlindBlack8[2], mid= 'white', limits= c(min_beta, max_beta), guide= 'none', midpoint= 0) +
scale_fill_gradientn(colours=c(colorBlindBlack8[4], 'white', colorBlindBlack8[2]), values= rescale(c(min_beta, 0, max_beta)), limits= c(min_beta, max_beta), guide= 'none') +
  coord_equal() +
  scale_x_discrete(labels= labs) +
  theme(axis.title= element_blank(),
        axis.ticks= element_blank(),
        plot.margin = margin(0, 0, 0, 0, "mm"),
        text= element_text(size= 9/ .pt),
        axis.text.y= element_text(hjust= 0.5),
	axis.text.x= element_text(angle= 45, hjust= 1),
        axis.line = element_line(colour = 'black', size = 0.2)) +
  geom_text_repel(data= filter(d, haplotype== 'Paternal\ntransmitted'), aes(x= rsid_lab, y= 4,
                label= round(probability, 2)),  direction= 'y', size= 8/ .pt, box.padding = 0.01)

ggsave(snakemake@output[[1]], plot= p1, width= 180, height= 60, units= 'mm', dpi= 300)

p1= ggplot(d, aes(rsid_lab, haplotype, fill= beta)) +
  theme_cowplot(8) +
  geom_tile() +
  scale_fill_gradientn(colours=c(colorBlindBlack8[4], 'white', colorBlindBlack8[2]), values= rescale(c(min_beta, 0, max_beta)), limits= c(min_beta, max_beta), name= 'Effect size') +
  coord_equal() +
scale_x_discrete(labels= labs) +  
theme(axis.title= element_blank(),
        axis.ticks= element_blank(),
        plot.margin = margin(0, 9, 0,0, "mm"),
        text= element_text(size= 9/ .pt),
        axis.text.y= element_text(hjust= 0.5),
        axis.line = element_line(colour = 'black', size = 0.2),
	legend.position= 'bottom') +
  geom_text_repel(data= filter(d, haplotype== 'Paternal\ntransmitted'), aes(x= rsid_lab, y= -0.05,
                                                                                label= round(probability, 2)), direction= "y" ,
                  size= 6.5/ .pt) 
ggsave(snakemake@output[[2]], plot= p1, width= 180, height= 100, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[3]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel gridExtra knitr scales plyr showtext ggtree dendextend From line 1 of figures/effect_origin_dendrogram.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(ggtern)
options(warn=-1)


x= fread(snakemake@input[[1]], h= T)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

x$rsid= with(x, ifelse(rsid== 'chrX:116013571', 'rs5991030', ifelse(rsid== 'chrX:132178061', 'rs5930554', rsid)))

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[2]], header= T, select= c('RSID', 'ID', 'nearestGene'))

x= inner_join(x, d, by= c('rsid'= 'RSID'))

x$nearestGene= with(x, ifelse(rsid== 'rs3129768', 'HLA-DQA1', ifelse(rsid== 'rs5991030', 'AGTR2', ifelse(rsid== 'rs5930554', 'RAP2C', nearestGene)))) 

x$nearestGene= with(x, ifelse(rsid== 'rs6780427', 'KCNAB1', nearestGene))

x$nearestGene= with(x, ifelse(rsid== 'rs6879092', 'EBF1', nearestGene))

#d= gather(x, haplotype, beta, c('Paternal', 'MaternalT', 'MaternalNT'))

#d$rsid_label= with(d, paste0(RSID, ' (', nearestGene, ')'))


#max_beta= max(abs(d$beta))

#d$class= factor(d$class, levels= c("MF SD", "MF OD", "Maternal", "Fetal MatT", "Fetal"))

#d= arrange(d, class, desc(max_prob))

#d$rsid_label= factor(d$rsid_label, levels= unique(d$rsid_label))


colT= colorBlindBlack8[4]
colR= colorBlindBlack8[1]
colL= colorBlindBlack8[2]

x$MF= x$MF_OD + x$MF_SD
x$Fet= x$Fetal_MatT + x$Fetal

p1= ggtern(x, aes(Maternal, Fet, MF)) +
geom_point(colour= colorBlindBlack8[8], fill= colorBlindBlack8[8], shape= 21) +
scale_alpha_continuous(range= c(0.6, 1), guide= F) +
scale_size_continuous(range= c(.001, 10), guide= F) +
theme_custom(tern.plot.background = NULL, tern.panel.background = 'white', col.T = colT, col.L = colL, col.R = colR, col.grid.minor = "white") +
Tarrowlab("Fetal only effect") +
Larrowlab("Maternal only effect") +
Rarrowlab("Maternal and fetal effect")  +
theme_showarrows()  +
theme_notitles() +
theme(text=element_text(family="arial", size= 10),
        tern.axis.arrow.T = element_blank(),
        tern.axis.arrow.L = element_blank(),
        tern.axis.arrow.R = element_blank(),
        tern.axis.text.T = element_text(color = colT),
        tern.axis.text.L = element_text(color = colL),
        tern.axis.text.R = element_text(color = colR),
        tern.axis.arrow.text.T = element_text(color = colT),
        plot.margin = margin(0, 0, 0, 0, "cm"),
        tern.axis.arrow.text.L = element_text(color = colL),
        tern.axis.arrow.text.R = element_text(color = colR),
        tern.panel.grid.major = element_line(linetype = 6, size = 0.3)) 
print('
ggtern(data=d, aes(-log10(pvalue_h1),-log10(pvalue_h2),-log10(pvalue_h3), label= nearestGene, size= abs(BETA), alpha= -log10(pvalue))) +
geom_point(colour= "black", fill= colorBlindBlack8[8], shape= 21) +
scale_alpha_continuous(range= c(0.6, 1), guide= F) +
scale_size_continuous(range= c(.001, 10), guide= F) +
theme_custom(tern.plot.background = NULL, tern.panel.background = "white", col.T = colT, col.L = colL, col.R = colR, col.grid.minor = "white") +
Tarrowlab("Maternal non-transmitted allele") + 
Larrowlab("Maternal transmitted allele") + 
Rarrowlab("Paternal transmitted allele")  +
theme_showarrows()  +
theme_notitles() +
theme(text=element_text(family="arial", size= 10),
	tern.axis.arrow.T = element_blank(),
	tern.axis.arrow.L = element_blank(),
	tern.axis.arrow.R = element_blank(),
        tern.axis.text.T = element_text(color = colT),
        tern.axis.text.L = element_text(color = colL),
        tern.axis.text.R = element_text(color = colR),
	tern.axis.arrow.text.T = element_text(color = colT), 
	plot.margin = margin(0, 0, 0, 0, "cm"), 
	tern.axis.arrow.text.L = element_text(color = colL),
	tern.axis.arrow.text.R = element_text(color = colR),
	tern.panel.grid.major = element_line(linetype = 6, size = 0.3)) +
geom_text(data= filter(d, nearestGene== "HAND2"), position= position_nudge_tern(y=0.05,x=-0.05/2,z=-0.05/2), aes(label=nearestGene), fontface= "bold", check_overlap=T, size= 8/ .pt, colour= "#525252", hjust= 1, vjust= 0.5)')


ggsave(snakemake@output[[1]], plot= p1, width= 95, height= 95, units= 'mm', dpi= 300)


d= select(x, rsid, ID, MF, Maternal, Fetal)

fwrite(x, snakemake@output[[2]], sep= '\t')
R dplyr data.table tidyr cowplot ggrepel knitr showtext ggtern From line 1 of figures/effect_origin_ternary.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
library(data.table)
library(dplyr)
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[1]])

d$lead_snp= with(d, ifelse(lead_snp== '1:50958027', '1:50959262', ifelse(lead_snp== '9:116929327', '9:116935764', ifelse(lead_snp== '5:157896786', '5:157895049', ifelse(lead_snp== '1:22511594', '1:22462111', lead_snp)))))

x= fread(snakemake@input[[2]])

x$lead_snp= paste(x$CHR, x$POS, sep= ':')

d= inner_join(d,x, by= 'lead_snp')

d$z_score= ifelse(d$z_score> 3.5, 3.5, d$z_score)

d$nearestGene= with(d, ifelse(nearestGene== 'CDC42', 'CDC42/ WNT4', ifelse(nearestGene== 'HIVEP3', 'HIVEP3/ EDN2', ifelse(nearestGene== 'TET3', 'TET3/ DGUOK-AS1', ifelse(nearestGene== 'TCEA2', 'TCEA2/ OPRL1', nearestGene)))))

d= filter(d, !(annotation %in% c('B2', 'geva_allele_age')))

d$annotation= with(d, ifelse(annotation== 'argweave', 'ARGWEAVE', 
		ifelse(annotation== 'betascore', 'Beta score',
		ifelse(annotation== 'B2', '', 
		ifelse(annotation== 'fst_eas_afr', 'Fst AFR-EAS',
		ifelse(annotation== 'fst_eur_afr', 'Fst AFR-EUR',
		ifelse(annotation== 'fst_eur_eas', 'Fst EAS-EUR',
		ifelse(annotation== 'gerp', 'GERP',
		ifelse(annotation== 'geva_allele_age', 'Alelle age',
		ifelse(annotation== 'iES_Sabeti', 'iES',
		ifelse(annotation== 'linsigh', 'LINSIGHT',
		ifelse(annotation== 'phastCon100', 'phastCONS100',
		ifelse(annotation== 'phyloP100', 'PhyloP',
		ifelse(annotation== 'xpehh_afr2_eas', 'XPEHH AFR-EAS',
		ifelse(annotation== 'xpehh_afr2_eur', 'XPEHH AFR-EUR',
		'XPEHH EAS-EUR')))))))))))))))

p1= ggplot(d, aes(annotation, nearestGene, fill= z_score)) +
geom_tile(colour = "white", size= 1) +
theme_cowplot(font_size= 9) +
scale_fill_gradient2(low= colorBlindBlack8[2], high= colorBlindBlack8[4], mid= 'white', limits= c(-2, 4)) +
theme(axis.text.x = element_text(angle = 45, hjust = 0),
        axis.title.x = element_blank(),
        axis.title.y = element_blank()) +
scale_x_discrete(position = "top") +
geom_text(data= filter(d, pvalue.x< 0.05), aes(annotation, nearestGene, label= '*'), size= 8/ .pt) +
theme(  panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.background = element_blank(),
        axis.ticks= element_blank(),
        panel.border = element_rect(colour= 'black', fill= NA, size=1),
        plot.margin = unit(c(0, 1, 0, 0), "cm"),
        axis.line= element_blank(),
	axis.text.y = element_text(face = "italic")) +
coord_equal()


ggsave(snakemake@output[[1]], plot= p1, width= 140, height= 120, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/evo.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


colorBlindBlack8  <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

d= fread(snakemake@input[[1]])

z= fread(snakemake@input[[3]])

df= fread(snakemake@input[[2]], select= (c('MarkerName', 'Effect', 'StdErr', 'HetISq', 'HetPVal', 'TOTALSAMPLESIZE', 'P-value', 'Allele1', 'Allele2')))
names(df)= c('SNP', 'BETA', 'SE', 'HetISq', 'HetPval', 'N', 'pvalue', 'A1', 'A2')
df= filter(df, SNP %in% d$SNP)

df= separate(df, SNP, into= c('CHR', 'POS', 'Ax1', 'Ax2', 'ID'), sep= ':', remove= F)
df$BETA= ifelse(df$A2 > df$A1, -1 * df$BETA, df$BETA)
df$CHR= ifelse(df$CHR== 'X','23', df$CHR)
df$CHR= as.integer(df$CHR)
df$POS= as.integer(df$POS)
df= select(df, -c(A1, A2, ID, Ax1, Ax2))

df$cohort= 'Meta-analysis'
d= bind_rows(d, df)

z$CHR= ifelse(z$CHR== 'X','23', z$CHR)
z$CHR= as.integer(z$CHR)

d= inner_join(d, z, by= 'CHR') %>% filter(POS> pos1, POS< pos2)

d$locus= paste0('Chr ', d$CHR,': ', d$nearestGene)

d= filter(d, !(cohort %in% c('PGPII', 'PGPIII', 'BIB', 'DNBCPTD', 'STORK', 'STORKGROR')))

d$cohort= paste0(d$cohort, ' (n= ', d$N, ')')

temp_df= d[d$nearestGene== snakemake@wildcards[['prev_locus']], ]

temp_df= temp_df[order(temp_df$N, decreasing= T), ]

rsid= ifelse(snakemake@wildcards[['prev_locus']]== 'EEFSEC', 'rs2659685', 
ifelse(snakemake@wildcards[['prev_locus']]== 'WNT4', 'rs12037376', 
ifelse(snakemake@wildcards[['prev_locus']]== 'EBF1', 'rs2963463',
ifelse(snakemake@wildcards[['prev_locus']]== 'AGTR2', 'rs5991030', 'rs28654158'))))

gene= unique(temp_df$nearestGene)
my_title = expression(paste0(italic(gene), " (,", rsid, ")"))

p1= ggplot(temp_df, aes(x=factor(cohort, level = factor(cohort)), y=BETA, ymin= BETA - 1.96 * SE, ymax= BETA + 1.96 * SE, colour= !is.na(HetISq), shape= !is.na(HetISq)), alpha= 0.5) +
 geom_pointrange(size= 0.4) +
scale_shape_manual(values= c(15, 18), guide= F) +
 geom_hline(yintercept = 0, linetype=2) +
scale_y_continuous(sec.axis = dup_axis()) +
ggtitle(parse(text = paste0(rsid, ' - ', "~italic('", unique(temp_df$nearestGene), "')"))) + 
coord_flip() +
scale_colour_manual(values= c(colorBlindBlack8[3], colorBlindBlack8[4]), guide= F) +
theme_cowplot(8) +
 xlab('') +
    ylab('Beta [95% CI]') +
geom_vline(xintercept= 0, linetype= "dotted", colour= 'grey') 

ggsave(snakemake@output[[1]], plot= p1, width= 140, height= 30.5  + 50/13 * nrow(temp_df), units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/forest_plot_EEFSEC.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])

d$term= with(d, ifelse(term== 'fetal_effect_PGS', 'Fetal', 'Maternal'))
d$outcome= gsub(' PGS', '', d$outcome)


p1= ggplot(d, aes(term, estimate, colour= term)) + 
geom_pointrange(aes(ymin= lo95, ymax= up95)) + 
facet_wrap(vars(outcome)) + 
scale_colour_manual(guide= 'none', values= colorBlindBlack8[c(2, 4)]) +
theme_cowplot(10) + 
geom_hline(yintercept= 0, colour= 'grey', size= 0.5, linetype= 'dashed') + 
theme(strip.background = element_blank(),
        panel.border = element_rect(colour = "black", fill = NA)) + 
ylab('Effect on gestational duration \ngenetic score (95% CI), days') +
xlab('Birth weight genetic score')


ggsave(snakemake@output[[1]], plot= p1, width= 180, height= 100, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/GA_BW_PGS_correlations.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
library(MendelianRandomization)
library(data.table)
library(dplyr)
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


top_ga= fread(snakemake@input[[1]])

top_ga= c(pull(top_ga, ID), '5:158058432:G:T', '3:156697097:A:G')


top_ptd= fread(snakemake@input[[2]])

top_ptd= pull(top_ptd, ID)

top= c(top_ga, top_ptd)

top= unique(top)

ga= fread(snakemake@input[[3]], select= c('ID', 'BETA', 'SE'))

ga= filter(ga, ID %in% top)

ptd= fread(snakemake@input[[4]], select= c('ID', 'BETA', 'SE'))

ptd= filter(ptd, ID %in% top) %>% select(ID, BETA, SE)

names(ptd)= c('ID', 'BETA_ptd', 'SE_ptd')

d= inner_join(ga, ptd, by= 'ID')

d$GWAS= with(d, ifelse(ID== '5:157895049:C:T', 'Both phenotypes', ifelse(ID %in% top_ptd, 'Preterm delivery', 'Gestational duration')))

p1= ggplot(d, aes(BETA, BETA_ptd, colour= GWAS, fill= GWAS)) +
geom_errorbarh(aes(xmin= BETA - SE, xmax= BETA + SE, colour= GWAS, fill= GWAS), size= 0.1, alpha= 0.7) +
geom_errorbar(aes(ymin= BETA_ptd - SE_ptd, ymax= BETA_ptd + SE_ptd, colour= GWAS, fill= GWAS),size= 0.1, alpha= 0.7) +
geom_point(size= 2, shape=21, stroke= 0.1, alpha= 0.7) +
scale_colour_manual(values= colorBlindBlack8[c(4, 2, 1)], guide= 'none') +
scale_fill_manual(values= colorBlindBlack8[c(4, 2, 1)], guide= 'none') +
xlab('Maternal effect on gestational duration, days') +
ylab('Maternal effect on preterm delivery, log(OR)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))


ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

p1= ggplot(d, aes(BETA, BETA_ptd, colour= GWAS, fill= GWAS)) +
geom_errorbarh(aes(xmin= BETA - SE, xmax= BETA + SE, colour= GWAS, fill= GWAS), size= 0.1, alpha= 0.7) +
geom_errorbar(aes(ymin= BETA_ptd - SE_ptd, ymax= BETA_ptd + SE_ptd, colour= GWAS, fill= GWAS),size= 0.1, alpha= 0.7) +
geom_point(size= 2, shape=21, stroke= 0.1, alpha= 0.7) +
scale_colour_manual(values= colorBlindBlack8[c(4, 2, 1)], guide= 'none') +
scale_fill_manual(values= colorBlindBlack8[c(4, 2, 1)]) +
xlab('Maternal effect on gestational duration, days') +
ylab('Maternal effect on preterm delivery, log(OR)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))

ggsave(snakemake@output[[2]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/GAraw_vs_allPTD.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)



pph= fread(snakemake@input[[1]])
supp_table= pph
geneb= fread(snakemake@input[[2]])

gene_dict= fread(snakemake@input[[3]])

names(gene_dict)= c('CHR', 'POS1', 'POS2', 'Gene', 'EnsembleID')

gene_dict$EID= with(gene_dict, unlist(lapply(strsplit(as.character(EnsembleID), ".", fixed= T), '[[', 1)))

d= inner_join(pph, gene_dict, by= c('protein'= 'EID')) %>% inner_join(., geneb, by= 'Gene')

#supp_table= full_join(pph, gene_dict, by= c('protein'= 'EID')) %>% full_join(., geneb, by= 'Gene') %>% filter(Pvalue< 0.05/ nrow(geneb) | PP.H4.abf>= 0.9)


z= fread(snakemake@input[[5]], select= c('z.df1', 'z.df2', 'SNP.PP.H4', 'protein', 'snp'))
z= arrange(z, desc(SNP.PP.H4))

z= group_by(z, protein) %>% filter(row_number()==1)

d= left_join(d, z, by= c('protein'))

d= separate(d, snp, into= c('CHR', 'POS', 'REF', 'EFF'), sep= ':', remove= FALSE)

#aa= fread(snakemake@input[[6]])
#names(aa)= c('CHR', 'POS', 'REF', 'ALT', 'AA')
#aa= filter(aa, AA!= '.')
#aa= filter(aa, POS %in% d$POS)

#aa$ID= with(aa, ifelse(REF> ALT, paste(CHR, POS, ALT, REF, sep= ':'), paste(CHR, POS, REF, ALT, sep= ':')))

#d= left_join(d,aa[, c('ID', 'AA')], by= c('snp'= 'ID'))

#d$z.df1= with(d, ifelse(d$AA== d$EFF, -1 * d$z.df1, d$z.df1))
#d$z.df2= with(d, ifelse(d$AA== d$EFF, -1 * d$z.df2, d$z.df2))

#d$direction= with(d, ifelse(z.df1>0 & z.df2 > 0, 'Positive', ifelse(z.df1<0 & z.df2< 0, 'Negative', 'Opposite')))
#d$direction= with(d, ifelse(is.na(d$AA), 'Missing', d$direction))

d$direction= with(d, ifelse((z.df1 * z.df2)>0, 'Same direction', 'Opposite'))

d$gene_group= with(d, ifelse(PP.H4.abf> 0.9 & Pvalue< 0.05 / nrow(geneb), 'Colocalize and gene-based significant', ifelse(Pvalue< 0.05 / nrow(geneb) & PP.H4.abf<= 0.9, 'Gene based significant',
	ifelse(PP.H4.abf> 0.9 & Pvalue> 0.05 / nrow(geneb), 'Colocalize', 'No colocalize and not significant'))))

ga= fread(snakemake@input[[4]], select= c('ID', 'BETA'))

d= inner_join(d, ga, by= c('snp'= 'ID'))

p1= ggplot(d, aes(-log10(Pvalue), PP.H4.abf, fill= direction, alpha= (1 + PP.H4.abf) * -log10(Pvalue))) +
geom_point(shape=21, colour= 'black', size= 4) +
theme_cowplot(font_size= 10) +
scale_alpha_continuous(guide= F) +
scale_size_continuous(range = c(.001, 10), guide= F) +
scale_fill_manual(values= c(colorBlindBlack8[c(2, 4)]), guide= F) +
geom_text_repel(data= filter(d, PP.H4.abf> 0.9 | Pvalue< 0.05 / nrow(geneb)), aes(label= Gene), max.overlaps= 20, colour= 'black', size= 6/ .pt, max.time= 10, alpha= 1) +
geom_hline(yintercept= 0.9, colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_vline(xintercept= -log10(0.05/nrow(geneb)), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
scale_y_continuous(breaks= c(seq(0, 1, 0.25), 0.9), limits= c(0, 1), expand= expansion(mult= c(0.05,0))) +
ylab('Posterior probability of colocalization') +
xlab('-log10(Gene based p-value)')

ggsave(snakemake@output[[1]], plot= p1, width= 95, height= 95, units= 'mm', dpi= 300)

d= select(d, Gene, BETA, direction, Pvalue, PP.H4.abf, Pvalue, z.df1, z.df2)

fwrite(d, snakemake@output[[2]], sep= '\t')

p1= ggplot(d, aes(-log10(Pvalue), PP.H4.abf, fill= direction, alpha= (1 + PP.H4.abf) * -log10(Pvalue))) +
geom_point(shape=21, colour= 'black', size= 4) +
theme_cowplot(font_size= 10) +
scale_alpha_continuous('Legend') +
scale_size_continuous('Legend', range = c(.001, 10)) +
scale_fill_manual('Legend', values= c(colorBlindBlack8[c(2, 4)])) +
geom_text_repel(data= filter(d, PP.H4.abf> 0.9 | Pvalue< 0.05 / nrow(geneb)), aes(label= Gene), max.overlaps= 20, colour= 'black', size= 6/ .pt, max.time= 10, alpha= 1) +
geom_hline(yintercept= 0.9, colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_vline(xintercept= -log10(0.05/nrow(geneb)), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
scale_y_continuous(breaks= c(seq(0, 1, 0.25), 0.9), limits= c(0, 1), expand= expansion(mult= c(0.05,0))) +
ylab('Posterior probability of colocalization') +
xlab('-log10(Gene based p-value)')

ggsave(snakemake@output[[3]], plot= p1, width= 90, height= 90, units= 'mm', dpi= 300)

fwrite(supp_table, snakemake@output[[4]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/gene_based_vs_coloc_iPSC.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
library(scales)
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(tidyverse)
library(fmsb)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

d= fread(snakemake@input[[1]])

d$p1= 'Gestational\nduration'
d$p2= with(d, ifelse(grepl('postTerm', p2), 'Post-term\ndelivery', ifelse(grepl('allPTD', p2), 'Preterm\ndelivery', 'GAnrm')))

d= filter(d, p2!= 'GAnrm')

p1= ggplot(d, aes(p2, rg, colour= p2)) +
  geom_point() +
geom_errorbar(aes(ymin= I(rg - 1.96*se) , ymax= (rg + 1.96 * se)), width=.2, position=position_dodge(.9)) +
theme_cowplot(font_size= 9) +
scale_fill_manual(values= colorBlindBlack8[c(8,3,2)], guide= 'none') +
scale_colour_manual(guide= 'none', values= colorBlindBlack8[c(8,3,2)]) +
xlab('Phenotype') +
ylab('Genetic correlation [95% CI]') +
theme(legend.position= 'none') +
ylim(pmin(-1, min(d$rg - 1.96*d$se)), pmax(1, max(d$rg + 1.96 * d$se))) +
geom_hline(yintercept= 0, linetype= 'dashed', colour= 'grey', size= 0.5)


ggsave(snakemake@output[[1]], plot= p1, width= 60, height= 80, units= 'mm', dpi= 300)
R ggplot2 tidyverse dplyr data.table tidyr cowplot ggrepel knitr scales showtext From line 1 of figures/genet_correlations_meta.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])

d= filter(d, pheno!= 'GAnrm')

d$pheno= with(d, ifelse(pheno== 'GAraw', 'Gestational\nduration', ifelse(pheno== 'allPTD', 'Preterm\ndelivery', 'Post-term\ndelivery')))

p1= ggplot(d, aes(pheno, h2, colour= pheno)) +
  geom_point() +
geom_errorbar(aes(ymin= I(h2 - 1.96*se) , ymax= (h2 + 1.96 * se)), width=.2, position=position_dodge(.9)) +
theme_cowplot(font_size= 9) +
scale_fill_manual(values= colorBlindBlack8[c(8,3,2)], guide= 'none') +
scale_colour_manual(guide= 'none', values= colorBlindBlack8[c(8,3,2)]) +
xlab('Phenotype') +
ylab('Common SNP heritability [95% CI]') +
theme(legend.position= 'none',
	axis.text.x= element_text(angle= 45, hjust= 1))


ggsave(snakemake@output[[1]], plot= p1, width= 60, height= 80, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/h2_allphenos.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)
x= fread(snakemake@input[[2]], h= T)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)
d$trait= 'Gestational\nduration'
x$trait= 'Preterm delivery'

d= rbind(d, x)

p1= ggplot(d, aes(cohort, h2, colour= cohort)) +
  geom_point() +
geom_errorbar(aes(ymin= I(h2 - 1.96*se) , ymax= (h2 + 1.96 * se)), width=.2, position=position_dodge(.9)) +
theme_cowplot(font_size= 9) +
facet_wrap(vars(trait), ncol= 1) +
scale_fill_manual(values= colorBlindBlack8[c(8,3,2,6,7, 4, 1)], guide= 'none') +
scale_colour_manual(guide= 'none', values= colorBlindBlack8[c(8,3,2,6,7, 4, 1)]) +
xlab('Cohort') +
ylab('Common SNP heritability [95% CI]') +
theme(legend.position= 'none',
	strip.background = element_blank(),
	axis.text.x= element_text(angle= 45, hjust= 1))


ggsave(snakemake@output[[1]], plot= p1, width= 60, height= 120, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/h2_cohorts.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
library(dplyr)
library(data.table)
library(ggplot2)
library(cowplot)
library(ggrepel)
library(tidyr)
library(showtext)
colorBlindBlack8  <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])
names(d)[8]= 'phenocode'
mani= fread(snakemake@input[[2]])

trait_list= c('biomarkers', 'continuous', 'icd10')
mani= mani[mani$trait_type %in% trait_list, ]

mani= filter(mani, saige_heritability_EUR> 0.01)
mani= mani[order(mani$saige_heritability_EUR, decreasing= TRUE), ]
mani= mani[!duplicated(mani$phenocode), ]

mani$phenocode= paste(mani$trait_type, mani$phenocode, sep= '_')
mani= mani[, c('phenocode', 'description')]
mani= mani[!duplicated(mani$description), ]

d= inner_join(d, mani[, c('description', 'phenocode')], by= 'phenocode')
d$cohort= 'UKBB'

x= fread(snakemake@input[[3]])
names(x)[8]= 'phenocode'
mani= fread(snakemake@input[[4]])
mani= mani[, c('phenocode', 'name')]
names(mani)= c('phenocode', 'description')
mani= mani[!duplicated(mani$description), ]

x= inner_join(x, mani, by= 'phenocode')
x$cohort= 'FINNGEN'

d= rbind(d, x)
d= d[order(d$PP.H4.abf, decreasing= F), ]
d= filter(d, PP.H4.abf> 0.01, PP.H4.abf + PP.H3.abf> 0.75)

d$preg_trait= factor(d$preg_trait)
empty_bar <- 6
to_add <- data.frame( matrix(NA, empty_bar*nlevels(d$preg_trait), ncol(d)) )
colnames(to_add) <- colnames(d)
to_add$preg_trait <- rep(levels(d$preg_trait), each=empty_bar)
d <- rbind(d, to_add)
d <- d %>% arrange(preg_trait)


d$id= seq(1, nrow(d))

label_data= d
number_of_bar <- nrow(label_data)
angle <-  90 - 360 * (label_data$id-0.5) /number_of_bar
label_data$hjust<-ifelse( angle < -90, 1, 0)


label_data$angle<-ifelse(angle < -90, angle+180, angle)

#d$id= factor(d$id, levels= d$id[order(d$PP.H4.abf)])

base_data= d %>%
  group_by(preg_trait) %>%
  filter(is.na(PP.H4.abf)) %>%
  summarize(start=min(id), end=max(id) ) %>%
  rowwise() %>%
  mutate(title=mean(c(start, end)))

arc100= rep(1, 2)
arc75= rep(0.75, 2)
arc50= rep(0.50, 2)
arc25= rep(0.25, 2)

p1= ggplot(d, aes(as.factor(id), PP.H4.abf, fill= preg_trait, alpha= PP.H4.abf)) +
geom_bar(stat="identity", colour= NA) +
scale_alpha_continuous(range= c(0.4, 1), guide= F) +
geom_segment(data=base_data, aes(x = end, y = arc100, xend = start, yend = arc100), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc75, xend = start, yend = arc75), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc50, xend = start, yend = arc50), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  geom_segment(data=base_data, aes(x = end, y = arc25, xend = start, yend = arc25), colour = "grey", alpha=1, size=0.3 , inherit.aes = FALSE ) +
  annotate("text", x = ((base_data$end[1] + base_data$start[1]) / 2), y = c((0.25 + 0.05) , (0.50 + 0.05), (0.75 + 0.05) , (1 + 0.05)), label = c("0.25", "0.50", "0.75", "1") , color="grey", size=3 , angle=0, fontface="bold", hjust= 0.5) +
   annotate("text", x = ((base_data$end[2] + base_data$start[2]) / 2), y = c((0.25 + 0.05) , (0.50 + 0.05), (0.75 + 0.05) , (1 + 0.05) ), label = c("0.25", "0.50", "0.75", "1") , color="grey", size=3, angle=15, fontface="bold", hjust=0.5) +
ylim(-0.2, 2) + # Limits of the plot = very important. The negative value controls the size of the inner circle, the positive one is useful to add size over each bar
theme_cowplot() +
scale_fill_manual(values=colorBlindBlack8[c(2,4)], guide= F) +
scale_colour_manual(values=colorBlindBlack8[c(2,4)], guide= F) +
  theme(
    axis.text = element_blank(),
    axis.title = element_blank(),
    panel.grid = element_blank(),
    plot.margin = unit(rep(-2,4), "cm")    ) +
  coord_polar(start = 0) +
geom_text(data=filter(label_data, PP.H4.abf> 0.75), aes(x= factor(id), y=PP.H4.abf + 0.01, label=description, hjust=hjust), color="black", fontface="bold",alpha=0.6, size=2.5, angle= filter(label_data, PP.H4.abf> 0.750)$angle, inherit.aes = FALSE) +
theme(panel.grid = element_blank(),
axis.title = element_blank(),
axis.text = element_blank(),
axis.ticks = element_blank())

p1= save_plot(snakemake@output[[1]], p1, base_width= 8, base_height= 8)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel showtext From line 1 of figures/KCNAB1_pheWAS.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
library(data.table)
library(dplyr)
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)



d= fread(snakemake@input[[1]])

d$Category= factor(d$Category, levels= unique(d$Category))

p1= ggplot(d, aes(Enrichment, -log10(Enrichment_p))) + 
geom_point(aes(size= Enrichment_p< 0.05/ (nrow(d)-1)), shape= 21, stroke= 0.1, fill= colorBlindBlack8[4]) +
xlab('Heritability enrichment') +
ylab('-log10(P-value)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(panel.grid.major= element_line(colour= 'grey', size= 0.05),
legend.position = "none")  +
geom_text_repel(data= filter(d, Enrichment_p< 0.05), aes(Enrichment, -log10(Enrichment_p), label= Category), size= 8/.pt)


ggsave(snakemake@output[[1]], plot= p1, width= 120, height= 90, units= 'mm', dpi= 300)


p2= ggplot(d, aes(n_genes, -log10(Enrichment_p))) + 
geom_point(aes(size= Enrichment_p< 0.05/ (nrow(d)-1)), shape= 21, stroke= 0.1, fill= colorBlindBlack8[4]) +
xlab('Size of gene set') +
ylab('-log10(Enrichment)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(panel.grid.major= element_line(colour= 'grey', size= 0.05),
legend.position = "none") +
geom_text_repel(data= filter(d, Enrichment_p< 0.05), aes(n_genes, -log10(Enrichment_p), label= Category), size= 8/.pt)

ggsave(snakemake@output[[2]], plot= p2, width= 90, height= 90, units= 'mm', dpi= 300)

p3= ggplot(d, aes(n_genes, -log10(Enrichment_p))) + 
geom_point(aes(size= Enrichment_p< 0.05/ (nrow(d)-1)), shape= 21, stroke= 0.1, fill= colorBlindBlack8[4]) +
xlab('Size of gene set') +
ylab('-log10(Enrichment)') +
theme_cowplot(font_size= 8) +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(panel.grid.major= element_line(colour= 'grey', size= 0.05)) +
geom_text_repel(data= filter(d, Enrichment_p< 0.05), aes(n_genes, -log10(Enrichment_p), label= Category), size= 8/.pt)

ggsave(snakemake@output[[3]], plot= p3, width= 90, height= 90, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table cowplot ggrepel showtext From line 1 of figures/labor_deg.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')



colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")


font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[1]])

x= fread(snakemake@input[[2]], select= c('RSID', 'BETA'))


d= inner_join(d, x, by= c('rsid' = 'RSID'))

d$beta_MNT= with(d, ifelse(BETA< 0, -1 * beta_MNT,  beta_MNT))
d$beta_PT= with(d, ifelse(BETA< 0, -1 * beta_PT, beta_PT))
d$beta_MT= with(d, ifelse(BETA< 0, -1 * beta_MT, beta_MT))
d$BETA= with(d, ifelse(BETA<0, -1 * BETA, BETA))

d$lo95_MT= d$beta_MT - 1.96 * d$se_MT
d$up95_MT= d$beta_MT + 1.96 * d$se_MT

d$lo95_MNT= d$beta_MNT - 1.96 * d$se_MNT
d$up95_MNT= d$beta_MNT + 1.96 * d$se_MNT

d$lo95_PT= d$beta_PT - 1.96 * d$se_PT
d$up95_PT= d$beta_PT + 1.96 * d$se_PT

d$class_name= with(d, ifelse(class_name== 'MF SD', 'Maternal and fetal (same direction)', ifelse(class_name== 'Fetal MatT', 'Fetal effect, maternal transmitted only', ifelse(class_name== 'Maternal', 'Maternal', ifelse(class_name== 'Fetal', 'Fetal', ifelse(class_name== 'MF OD', 'Maternal and fetal (opposite direction)', ''))))))

p1= ggplot(d, aes(beta_MNT, BETA, colour= class_name)) +
geom_point(size= 0.5) +
#geom_errorbarh(data= filter(d, (lo95_h2 >0 & up95_h2>0) | (lo95_h2<0 & up95_h2 <0)), aes(xmax = lo95_h2, xmin = up95_h2), size= 0.05) +
theme_cowplot(font_size= 8) +
scale_colour_manual(values= c('grey', colorBlindBlack8[c(8, 2, 4, 3)])) +
geom_vline(xintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_hline(yintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
xlab('Effect size maternal \nnon-transmitted alleles, days') +
ylab('Effect size maternal genome, days')
#theme(legend.direction = "horizontal", legend.position = "bottom")
#scale_x_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1)) +
#  scale_y_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1))


ggsave(snakemake@output[[1]], plot= p1, width= 60, height= 60, units= 'mm', dpi= 300)

print('plot1')
p1= ggplot(d, aes(beta_PT, BETA, colour= class_name)) +
geom_point(size= 0.5) +
#geom_errorbarh(data= filter(d, (lo95_h3 >0 & up95_h3>0) | (lo95_h3<0 & up95_h3 <0)), aes(xmax = lo95_h3, xmin = up95_h3), size= 0.05) +
theme_cowplot(font_size= 8) +
scale_colour_manual(values= c('grey', colorBlindBlack8[c(8, 2, 4, 3)])) +
geom_vline(xintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_hline(yintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
xlab('Effect size paternal \ntransmitted alleles, days') +
ylab('Effect size maternal genome, days') 
#scale_x_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1)) +
#  scale_y_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1))

ggsave(snakemake@output[[2]], plot= p1, width= 60, height= 60, units= 'mm', dpi= 300)

print('plot2')
p1= ggplot(d, aes(beta_MT, BETA, colour= class_name)) +
geom_point(size= 0.5) +
#geom_errorbarh(data= filter(d, (lo95_h3 >0 & up95_h3>0) | (lo95_h3<0 & up95_h3 <0)), aes(xmax = lo95_h3, xmin = up95_h3), size= 0.05) +
theme_cowplot(font_size= 8) +
scale_colour_manual(values= c('grey', colorBlindBlack8[c(8, 2, 4, 3)]), guide= F) +
geom_vline(xintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_hline(yintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
xlab('Effect size maternal \ntransmitted alleles, days') +
ylab('Effect size maternal genome, days')
#scale_x_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1)) +
#  scale_y_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1))

ggsave(snakemake@output[[3]], plot= p1, width= 60, height= 60, units= 'mm', dpi= 300)

p1= ggplot(d, aes(beta_MNT, BETA, colour= class_name)) +
geom_point(size= 0.5) +
#geom_errorbarh(data= filter(d, (lo95_h2 >0 & up95_h2>0) | (lo95_h2<0 & up95_h2 <0)), aes(xmax = lo95_h2, xmin = up95_h2), size= 0.05) +
theme_cowplot(font_size= 8) +
scale_colour_manual(values= c('grey', colorBlindBlack8[c(8, 2, 4, 3)])) +
geom_vline(xintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_hline(yintercept= 0, colour= colorBlindBlack8[1], linetype= 'dashed', size= 0.2, alpha= 0.6) +
xlab('Effect size maternal \nnon-transmitted alleles, days') +
ylab('Effect size maternal genome, days') 
theme(legend.direction = "horizontal", legend.position = "bottom")
#scale_x_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1)) +
#  scale_y_continuous(breaks = round(seq(-1.5, 3, by= 0.5), 1))

ggsave(snakemake@output[[4]], plot= p1, width= 120, height= 60, units= 'mm', dpi= 300)
fwrite(d, snakemake@output[[5]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/lm_effect_origin.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])

x= fread(snakemake@input[[2]])

d= rbind(d, x)

names(d)= c('Name', 'no_no', 'no_yes', 'yes_no', 'yes_yes', 'candidate_gene', 'rest_genes', 'OR', 'pvalue')
d$enrichment= d$candidate_gene / d$rest_genes

d= arrange(d, desc(pvalue))

d$description= with(d, ifelse(Name== 'pli', 'Loss-of-function intolerant',
			ifelse(Name== 'dominant', 'Dominant', 'Recessive')))

d$description= factor(d$description, levels= unique(d$description))



p1= ggplot(data=d, aes(x= description, y= -log10(pvalue))) +
geom_col(fill=colorBlindBlack8[2], alpha= 0.6) +
theme_cowplot(font_size= 10) +
ylab('Enrichment -log10(pvalue)') +
theme(axis.title.y=element_blank()) +
geom_hline(yintercept= -log10(0.05/nrow(d)), linetype= 'dashed', colour= 'grey') +
coord_flip()


ggsave(snakemake@output[[1]], plot= p1, height= 35, width= 90, dpi= 300, units= 'mm')

fwrite(d, snakemake@output[[2]], sep='\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/MacArthurlab_enrichment.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

ge= data.frame(CHR= c(5, 3, 1, 23, 3, 23), pos_ge= c(157895049, 127881613, 22470407, 115164770, 123068359, 131300571))

df= arrange(d, pvalue)

dg= fread(snakemake@input[[2]])
dg$GENE= dg$nearestGene


don <- df %>%
    group_by(CHR)      %>%
    summarise(chr_len= max(POS)) %>%
    mutate(tot= cumsum(as.numeric(chr_len))-chr_len) %>% # Calculate cumulative position of each chromosome
    select(-chr_len) %>%
    left_join(df, ., by= 'CHR') %>%
    arrange(CHR, POS) %>% # Add a cumulative position of each SNP
    mutate(BPcum=POS+tot) %>%
         ungroup()

axisdf = don %>% group_by(CHR) %>% summarize(center=( max(BPcum) + min(BPcum) ) / 2 )
  names(axisdf)= c('CHR', 'center')
HC= -log10(5*10**-8)
dg= dg %>% ungroup() %>% select(ID, GENE, CHR, POS, BETA)

don$disc= ifelse(don$pvalue> 5*10**-8, 0, 2)

don= left_join(don, select(dg, ID, GENE), by= 'ID')
names(dg)= c('ID', 'GENE', 'CHR', 'POS_new', 'BETA')

lims= 250000

don= data.frame(don)
dg= data.frame(dg)


for (i in rownames(dg)) {
don= mutate(don, disc= ifelse(CHR== as.integer(dg[i, 'CHR']) & POS>= as.integer(dg[i, 'POS_new']) - lims & POS<= as.integer(dg[i, 'POS_new']) + lims, 2, disc)) 
}


for (i in rownames(ge)) {
don= mutate(don, disc= ifelse(CHR== as.integer(ge[i, 'CHR']) & POS>= as.integer(ge[i, 'pos_ge']) - lims & POS<= as.integer(ge[i, 'pos_ge']) + lims, 1, disc))
}

don= don[order(don$disc, decreasing= F, na.last= T), ]
don$disc= factor(don$disc, levels=c(0, 1, 2), labels=c('Not significant', 'Previous discovery', 'New discovery'))

cols <- c('Not significant'= 'grey', 'Previous discovery'= colorBlindBlack8[3], 'New discovery'= colorBlindBlack8[8])

don$GENE= ifelse(!is.na(don$GENE), don$nearestGene, don$GENE)

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


don$GENE= with(don, ifelse(GENE== 'CDC42', 'CDC42/ WNT4', ifelse(GENE== 'HIVEP3', 'HIVEP3/ EDN2', ifelse(GENE== 'TET3', 'TET3/ DGUOK-AS1', ifelse(GENE== 'TCEA2', 'TCEA2/ OPRL1', GENE)))))

don$logpval= -log10(don$pvalue)

p1= ggplot(data= don, aes(x= BPcum, y= logpval, colour= disc)) +
  geom_point(size= 0.07) +   # Show all points
  theme_cowplot(font_size= 9) +
  scale_colour_manual(values= cols, guide= F) +
  scale_x_continuous(label = c(1:19, '', 21,'', 'X'), breaks= axisdf$center, expand= c(0.03, 0.03)) + # label = ifelse(axisdf$CHR== 23, 'X', axisdf$CHR)
  scale_y_continuous(expand= c(0, 0), limits= c(min(don$logpval) - 2, max(don$logpval) + 2), breaks= seq(0, 10, 5), labels= c(abs(seq(0, 10, 5)))) + # , sec.axis = sec_axis(~ ., name = derive())) +
  ylab('-log10(pvalue)') +
  xlab('Chromosome') +
  geom_hline(yintercept= 0,, size= 0.25, colour= 'black') +
  geom_hline(yintercept= c(HC, -HC), size= 0.2, linetype= 2, colour= '#878787') +
  coord_cartesian(clip = "off") +
  geom_text_repel(data= filter(don, GENE!= ''), aes(x= BPcum, y= logpval, label= GENE),
                  size= 6/ .pt,
                  force_pull= 0, # do not pull toward data points
                  force= 0.1,
                  nudge_y      =  ifelse(filter(don, GENE!= '') %>% pull(logpval)>0, 1, -1), #43 - ((-log10(filter(don, GENE!= '')$pvalue))),
                  direction    = "both",
                  hjust        = 0,
                  vjust=  0.5,
		  box.padding= 0.1,
		  angle= 0,
                  segment.size = 0.1,
                  segment.square= TRUE,
                  segment.inflect= FALSE,
                  segment.colour= colorBlindBlack8[8],
                  colour= ifelse(filter(don, GENE!= '') %>% pull(disc)== 'New discovery', colorBlindBlack8[8], colorBlindBlack8[3]),
                  segment.linetype = 4,
                  ylim = c(-Inf, 50),
                  xlim = c(-Inf, Inf)) +
  theme(legend.position= 'none',
	plot.margin = unit(c(t= 0, r=0, b= 0, l=0), 'cm'),
        text= element_text(family="arial", size= 9),
	axis.line= element_line(size= 0.1)) 

save_plot(snakemake@output[[1]], plot= p1, base_height= 90, base_width= 185, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/manhattan_plot_postTerm.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T, select= c('ID', 'CHR', 'POS', 'pvalue', 'nearestGene'))
d$pheno= 'GAraw'
x= fread(snakemake@input[[3]], h= T, select= c('ID', 'CHR', 'POS', 'pvalue', 'nearestGene'))
x$pheno= 'allPTD'

d= rbind(d, x)

rm(x)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

ge= data.frame(CHR= c(5, 3, 1, 23, 3, 23), pos_ge= c(157895049, 127881613, 22470407, 115164770, 123068359, 131300571))

df= arrange(d, pvalue)


dg= fread(snakemake@input[[2]])
dg$GENE= dg$nearestGene

ptd= fread(snakemake@input[[4]])
ptd$GENE= ptd$nearestGene

don <- df %>%
    group_by(CHR)      %>%
    summarise(chr_len= max(POS)) %>%
    mutate(tot= cumsum(as.numeric(chr_len))-chr_len) %>% # Calculate cumulative position of each chromosome
    select(-chr_len) %>%
    left_join(df, ., by= 'CHR') %>%
    arrange(CHR, POS) %>% # Add a cumulative position of each SNP
    mutate(BPcum=POS+tot) %>%
         ungroup()

axisdf = don %>% group_by(CHR) %>% summarize(center=( max(BPcum) + min(BPcum) ) / 2 )
  names(axisdf)= c('CHR', 'center')
HC= -log10(5*10**-8)
dg= dg %>% ungroup() %>% select(ID, GENE, CHR, POS, BETA)
ptd= ptd %>% ungroup %>% select(ID, GENE, CHR, POS, BETA)

don$disc= ifelse(don$pvalue> 5*10**-8, 0, 2)

don1= filter(don, pheno== 'GAraw') %>% left_join(., select(dg, ID, GENE), by= 'ID')
don2= filter(don, pheno!= 'GAraw') %>% left_join(., select(ptd, ID, GENE), by= 'ID')
names(dg)= c('ID', 'GENE', 'CHR', 'POS_new', 'BETA')
names(ptd)= c('ID', 'GENE', 'CHR', 'POS_new', 'BETA')

lims= 250000

don= data.frame(don)
dg= data.frame(dg)
ptd= data.frame(ptd)


for (i in rownames(dg)) {
don1= mutate(don1, disc= ifelse(CHR== as.integer(dg[i, 'CHR']) & POS>= as.integer(dg[i, 'POS_new']) - lims & POS<= as.integer(dg[i, 'POS_new']) + lims, 2, disc)) 
}

for (i in rownames(ptd)) {
don2= mutate(don2, disc= ifelse(CHR== as.integer(ptd[i, 'CHR']) & POS>= as.integer(ptd[i, 'POS_new']) - lims & POS<= as.integer(ptd[i, 'POS_new']) + lims, 2, disc))

}

don= rbind(don1, don2)
rm(don1) ; rm(don2)

for (i in rownames(ge)) {
don= mutate(don, disc= ifelse(CHR== as.integer(ge[i, 'CHR']) & POS>= as.integer(ge[i, 'pos_ge']) - lims & POS<= as.integer(ge[i, 'pos_ge']) + lims, 1, disc))
}

don= don[order(don$disc, decreasing= F, na.last= T), ]
don$disc= factor(don$disc, levels=c(0, 1, 2), labels=c('Not significant', 'Previous discovery', 'New discovery'))

cols <- c('Not significant'= 'grey', 'Previous discovery'= colorBlindBlack8[3], 'New discovery'= colorBlindBlack8[8])

don$GENE= ifelse(!is.na(don$GENE), don$nearestGene, don$GENE)

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


don$GENE= with(don, ifelse(GENE== 'CDC42', 'CDC42/ WNT4', ifelse(GENE== 'HIVEP3', 'HIVEP3/ EDN2', ifelse(GENE== 'TET3', 'TET3/ DGUOK-AS1', ifelse(GENE== 'TCEA2', 'TCEA2/ OPRL1', GENE)))))

don$logpval= with(don, ifelse(pheno== 'allPTD', log10(pvalue), -log10(pvalue)))

p1= ggplot(data= don, aes(x= BPcum, y= logpval, colour= disc)) +
  geom_point(size= 0.07) +   # Show all points
  theme_cowplot(font_size= 9) +
  scale_colour_manual(values= cols, guide= F) +
  scale_x_continuous(label = c(1:19, '', 21,'', 'X'), breaks= axisdf$center, expand= c(0.03, 0.03)) + # label = ifelse(axisdf$CHR== 23, 'X', axisdf$CHR)
  scale_y_continuous(expand= c(0, 0), limits= c(min(don$logpval) - 2, max(don$logpval) + 2), breaks= seq(-30, 45, 10), labels= c(abs(seq(-30, 45, 10)))) + # , sec.axis = sec_axis(~ ., name = derive())) +
  ylab('-log10(pvalue)') +
  xlab('Chromosome') +
  geom_hline(yintercept= 0,, size= 0.25, colour= 'black') +
  geom_hline(yintercept= c(HC, -HC), size= 0.2, linetype= 2, colour= '#878787') +
  coord_cartesian(clip = "off") +
  geom_text_repel(data= filter(don, GENE!= ''), aes(x= BPcum, y= logpval, label= GENE),
                  size= 6/ .pt,
                  force_pull= 0, # do not pull toward data points
                  force= 0.1,
                  nudge_y      =  ifelse(filter(don, GENE!= '') %>% pull(logpval)>0, 1, -1), #43 - ((-log10(filter(don, GENE!= '')$pvalue))),
                  direction    = "both",
                  hjust        = 0,
                  vjust=  0.5,
		  box.padding= 0.1,
		  angle= 0,
                  segment.size = 0.1,
                  segment.square= TRUE,
                  segment.inflect= FALSE,
                  segment.colour= colorBlindBlack8[8],
                  colour= ifelse(filter(don, GENE!= '') %>% pull(disc)== 'New discovery', colorBlindBlack8[8], colorBlindBlack8[3]),
                  segment.linetype = 4,
                  ylim = c(-Inf, 50),
                  xlim = c(-Inf, Inf)) +
  theme(legend.position= 'none',
	plot.margin = unit(c(t= 0, r=0, b= 0, l=0), 'cm'),
        text= element_text(family="arial", size= 9),
	axis.line= element_line(size= 0.1)) 

save_plot(snakemake@output[[1]], plot= p1, base_height= 90, base_width= 180, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/manhattan_plot.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

as= 8
as1= 8

d= fread(snakemake@input[[1]])
d$effect= 'fetal_effect'
x= fread(snakemake@input[[2]])
x$effect= 'maternal_effect'

d= rbind(d, x)

d= filter(d, !(rsID %in% c('rs7819593', 'rs41311445')))
d$Beta2= ifelse(d$Beta1< 0, -1 * d$Beta2, d$Beta2)
d$Beta1= ifelse(d$Beta1< 0, -1 * d$Beta1, d$Beta1)

d$beta_dif= with(d, (Beta2 - Beta1) / Beta1)

mor= filter(d, effect == 'maternal_effect') %>% pull(beta_dif)
barn= filter(d, effect == 'fetal_effect') %>% pull(beta_dif)

p1= ggplot() +
geom_density( mapping=aes(x = mor, y = ..density..), fill= colorBlindBlack8[3], colour= colorBlindBlack8[3]) +
annotate('text', x= 0.35, y= 0.6, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x= 0.35, y= -1, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
geom_density(mapping= aes(x = barn, y = -..density..), fill= colorBlindBlack8[8], colour= colorBlindBlack8[8]) +
  theme_cowplot(font_size = 8) +
scale_x_continuous(expand= c(0, 0)) +
  xlab("Relative difference in effect size on \nbirth weight with or without adjusting for gestational duration") +
ylab('Density') +
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)



p1= ggplot(d, aes(beta_dif, group= effect, fill= effect)) +
geom_hline(yintercept= 0, colour= 'black') +
geom_density(color= NA) +
annotate('text', x=-1.5, y= 0.8, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=1, y= 0.8, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
theme_cowplot(font_size= 8) +
#scale_colour_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(expand=c(0, 0.05)) +
  xlab("Relative difference in effect size on birth weight\nwith or without adjusting for gestational duration") +
ylab('Density') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3)) +
geom_vline(xintercept= 0, linetpye= 'dashed', colour= 'grey')

ggsave(snakemake@output[[3]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/mediation_BW_GA_individual_level_data_decode.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

as= 8
as1= 8

d= fread(snakemake@input[[1]])

d$beta_h2_GA= ifelse(d$beta_h2< 0, -1 * d$beta_h2_GA, d$beta_h2_GA)
d$beta_h2= ifelse(d$beta_h2< 0, -1 * d$beta_h2, d$beta_h2)

d$beta_h3_GA= ifelse(d$beta_h3 < 0, -1 * d$beta_h3_GA, d$beta_h3_GA)
d$beta_h3= ifelse(d$beta_h3 < 0, -1 * d$beta_h3, d$beta_h3)

d$beta_dif_h2= with(d, (beta_h2_GA - beta_h2) / beta_h2)
d$beta_dif_h3= with(d, (beta_h3_GA - beta_h3) / beta_h3)

mor= filter(d, effect == 'maternal_effect') %>% pull(beta_dif_h2)
barn= filter(d, effect == 'fetal_effect') %>% pull(beta_dif_h3)

p1= ggplot() +
geom_density( mapping=aes(x = mor, y = ..density..), fill= colorBlindBlack8[3], colour= colorBlindBlack8[3]) +
annotate('text', x= 0.35, y= 0.6, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x= 0.35, y= -1, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
geom_density(mapping= aes(x = barn, y = -..density..), fill= colorBlindBlack8[8], colour= colorBlindBlack8[8]) +
  theme_cowplot(font_size = 8) +
scale_x_continuous(expand= c(0, 0)) +
  xlab("Relative difference in effect size on \nbirth weight with or without adjusting for gestational duration") +
ylab('Density') +
geom_hline(yintercept= 0, colour= 'grey') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3))


ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)



moms= filter(d, effect== 'maternal_effect') %>% gather(key, beta_dif, beta_dif_h2) %>% select(beta_dif, effect)
fets= filter(d, effect== 'fetal_effect') %>% gather(key, beta_dif, beta_dif_h3) %>% select(beta_dif, effect)

d= rbind(moms, fets)


p1= ggplot(d, aes(beta_dif, group= effect, fill= effect)) +
geom_hline(yintercept= 0, colour= 'black') +
geom_density(color= NA) +
annotate('text', x=-2, y= 0.4, label= "Maternal", color= colorBlindBlack8[3], size= as1/ .pt, fontface = 'bold') +
annotate('text', x=1, y= 0.8, label="Fetal", color= colorBlindBlack8[8], size= as1/ .pt, fontface = 'bold') +
theme_cowplot(font_size= 8) +
#scale_colour_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(8,3)], 0.5), guide= 'none') +
scale_x_continuous(expand= c(0, 0)) +
scale_y_continuous(expand=c(0, 0.05)) +
  xlab("Relative difference in effect size on birth weight\nwith or without adjusting for gestational duration") +
ylab('Density') +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3)) +
geom_vline(xintercept= 0, linetpye= 'dashed', colour= 'grey')

ggsave(snakemake@output[[3]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/mediation_BW_GA_individual_level_data.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
library(MendelianRandomization)
library(data.table)
library(dplyr)
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])

d$rsid= with(d, ifelse(rsid== 'chrX:116013571', 'rs5991030', ifelse(rsid== 'chrX:132178061', 'rs5930554', rsid)))

d$effect_origin= with(d, ifelse(class_name== 'MF OD' | class_name== 'MF SD', 'Maternal and fetal', ifelse(class_name== 'Fetal MatT' | class_name== 'Fetal', 'Fetal', 'Maternal')))

#d= filter(d, MarkerName!= '6:32595083:G:T')

#top= fread(snakemake@input[[2]])
#ids= pull(top, ID)
#ids= c('3:156697097:A:G', '5:158058432:G:T', ids)

x= fread(snakemake@input[[2]], select= c('ID', 'RSID'))

#x= filter(x, ID %in% ids)

d= inner_join(d, x, by= c('rsid'= 'RSID'))

d= separate(d, ID, into= c('CHR', 'POS', 'REF', 'EFF'), sep= ':')
d$beta_MT= with(d, ifelse(REF > EFF, -1 * beta_MT, beta_MT))
d$beta_MNT= with(d, ifelse(REF > EFF, -1 * beta_MNT, beta_MNT))
d$beta_PT= with(d, ifelse(REF > EFF, -1 * beta_PT, beta_PT))

d$ID= with(d, ifelse(REF> EFF, paste(CHR, POS, EFF, REF, sep= ':'), paste(CHR, POS, REF, EFF, sep= ':')))

outcome= ifelse(grepl('fetal', snakemake@input[[3]]), 'Fetal', 'Maternal')

x= fread(snakemake@input[[3]], select= c('ID', 'BETA', 'SE', 'pvalue'))

d= inner_join(d, x, by= 'ID')

df_MT= select(d, beta_MT, se_MT, BETA, SE, effect_origin)
df_MT$BETA= with(df_MT, ifelse(beta_MT<0, BETA * -1, BETA))
df_MT$beta_MT= with(df_MT, ifelse(beta_MT<0, beta_MT * -1, beta_MT))


inputMR_m= mr_input(bx= df_MT$beta_MT, bxse= df_MT$se_MT, by= df_MT$BETA, byse= df_MT$SE)
MT= mr_allmethods(inputMR_m)$Values
names(MT)= c('method', 'estimate', 'se', 'lo95', 'up95', 'pvalue')

df_MNT= select(d, beta_MNT, se_MNT, BETA, SE, effect_origin)
df_MNT$BETA= with(df_MNT, ifelse(beta_MNT<0, BETA * -1, BETA))
df_MNT$beta_MNT= with(df_MNT, ifelse(beta_MNT<0, beta_MNT * -1, beta_MNT))


inputMR_m= mr_input(bx= df_MNT$beta_MNT, bxse= df_MNT$se_MNT, by= df_MNT$BETA, byse= df_MNT$SE)
MNT= mr_allmethods(inputMR_m)$Values
names(MNT)= c('method', 'estimate', 'se', 'lo95', 'up95', 'pvalue')


df_PT= select(d, beta_PT, se_PT, BETA, SE, effect_origin)
print(nrow(df_PT))
df_PT$BETA= with(df_PT, ifelse(beta_PT<0, BETA * -1, BETA))
df_PT$beta_PT= with(df_PT, ifelse(beta_PT<0, beta_PT * -1, beta_PT))

inputMR_m= mr_input(bx= df_PT$beta_PT, bxse= df_PT$se_PT, by= df_PT$BETA, byse= df_PT$SE)
PT= mr_allmethods(inputMR_m)$Values
names(PT)= c('method', 'estimate', 'se', 'lo95', 'up95', 'pvalue')

p1= ggplot(df_MT, aes(beta_MT, BETA, colour= effect_origin, fill= effect_origin)) +
geom_errorbarh(aes(xmin= beta_MT - se_MT, xmax= beta_MT + se_MT, colour= effect_origin, fill= effect_origin), size= 0.1, alpha= 0.7) +
geom_errorbar(aes(ymin= BETA - SE, ymax= BETA + SE, colour= effect_origin, fill= effect_origin),size= 0.1, alpha= 0.7) +
geom_point(size= 2, shape=21, stroke= 0.1, alpha= 0.7) +
scale_colour_manual(values= colorBlindBlack8[c(4, 2, 1)], guide= 'none') +
scale_fill_manual(values= colorBlindBlack8[c(4, 2, 1)], guide= 'none') +
xlab('Effect of maternal transmitted\nalleles on gestational duration, days') +
ylab(paste(outcome, 'only effect\non birth weight, z-score')) +
theme_cowplot(font_size= 8) +
geom_abline(intercept= 0, slope= filter(MT, method== 'IVW') %>% pull(estimate), colour= '#d9d9d9') +
geom_abline(intercept= (filter(MT, method== '(intercept)') %>% pull(estimate))[1], slope= filter(MT, method== 'MR-Egger') %>% pull(estimate), colour= '#d9d9d9', linetype= 'dashed') +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))

p2= ggplot(df_MNT, aes(beta_MNT, BETA, colour= effect_origin, fill= effect_origin)) +
geom_errorbarh(aes(xmin= beta_MNT - se_MNT, xmax= beta_MNT + se_MNT,colour= effect_origin, fill= effect_origin), size= 0.1) +
geom_errorbar(aes(ymin= BETA - SE, ymax= BETA + SE,colour= effect_origin, fill= effect_origin),size= 0.1) +
geom_point(size= 2, shape= 21, stroke= 0.1) +
scale_colour_manual(values= alpha(colorBlindBlack8[c(4, 2, 1)], 0.7), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(4, 2, 1)], 0.7), guide= 'none') +
xlab('Effect of maternal non-transmitted alleles\non gestational duration, days') +
ylab(paste(outcome, 'only effect\non birth weight, z-score')) +
theme_cowplot(font_size= 8) +
geom_abline(intercept= 0, slope= filter(MNT, method== 'IVW') %>% pull(estimate), colour= '#d9d9d9') +
geom_abline(intercept= (filter(MNT, method== '(intercept)') %>% pull(estimate))[1], slope= filter(MNT, method== 'MR-Egger') %>% pull(estimate), colour= '#d9d9d9', linetype= 'dashed') +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))

p3= ggplot(df_PT, aes(beta_PT, BETA, colour= effect_origin, fill= effect_origin)) +
geom_errorbarh(aes(xmin= beta_PT - se_PT, xmax= beta_PT + se_PT, colour= effect_origin, fill= effect_origin), size= 0.1) +
geom_errorbar(aes(ymin= BETA - SE, ymax= BETA + SE, colour= effect_origin, fill= effect_origin), alpha= 0.5, size= 0.1) +
geom_point(size= 2, shape= 21, stroke = 0.1) +
scale_colour_manual(values= alpha(colorBlindBlack8[c(4, 2, 1)], 0.7), guide= 'none') +
scale_fill_manual(values= alpha(colorBlindBlack8[c(4, 2, 1)], 0.7), guide= 'none') +
xlab('Effect of paternal transmitted alleles\non gestational duration, days') +
ylab(paste(outcome, 'only effect\non birth weight, z-score')) +
theme_cowplot(font_size= 8) +
geom_abline(intercept= 0, slope= filter(PT, method== 'IVW') %>% pull(estimate), colour= '#d9d9d9') +
geom_abline(intercept= (filter(PT, method== '(intercept)') %>% pull(estimate))[1], slope= filter(PT, method== 'MR-Egger') %>% pull(estimate), colour= '#d9d9d9', linetype= 'dashed') +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))

ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)
ggsave(snakemake@output[[2]], plot= p2, width= 70, height= 70, units= 'mm', dpi= 300)
ggsave(snakemake@output[[3]], plot= p3, width= 70, height= 70, units= 'mm', dpi= 300)

MT$haplotype= 'MT'
MNT$haplotype= 'MNT'
PT$haplotype= 'PT'

df= bind_rows(MT, MNT, PT)

fwrite(d, snakemake@output[[4]], sep= '\t')
fwrite(df, snakemake@output[[5]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/MR_GA_BW_haplotype.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
library(MendelianRandomization)
library(data.table)
library(dplyr)
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


d= fread(snakemake@input[[1]])
x= fread(snakemake@input[[2]], select= c('ID', 'BETA', 'SE'))

mr= fread(snakemake@input[[3]])

d= inner_join(d,x, by= 'ID')
d= filter(d, !duplicated(ID))

d$BETA= with(d, ifelse(beta< 0, -1 * BETA, BETA))
d$beta= with(d, ifelse(beta< 0, -1 * beta, beta))


shbg= filter(d, trait== 'SHBG_fem_cluster')
testo= filter(d, trait== 'Testosterone_fem_cluster')

p1= ggplot(shbg, aes(beta, BETA), color= colorBlindBlack8[2]) +
geom_errorbarh(aes(xmin= beta - se, xmax= beta + se), size= 0.1, alpha= 0.7, color= colorBlindBlack8[2]) +
geom_errorbar(aes(ymin= BETA - SE, ymax= BETA + SE), size= 0.1, alpha= 0.7, color= colorBlindBlack8[2]) +
geom_point(size= 2, shape=21, stroke= 0.1, alpha= 0.7, fill= colorBlindBlack8[2]) +
xlab('Effect on SHBG (women), nmol/L') +
ylab('Effect on gestational duration, days') +
theme_cowplot(font_size= 8) +
geom_abline(intercept= 0, slope= filter(mr, method== 'IVW', trait== 'SHBG_fem_cluster') %>% pull(estimate), colour= '#d9d9d9') +
geom_abline(intercept= (filter(mr, method== '(intercept)', trait== 'SHBG_fem_cluster') %>% pull(estimate))[1], slope= filter(mr, method== 'MR-Egger', trait== 'SHBG_fem_cluster') %>% pull(estimate), colour= '#d9d9d9', linetype= 'dashed') +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))


p2= ggplot(testo, aes(beta, BETA), color= colorBlindBlack8[2]) +
geom_errorbarh(aes(xmin= beta - se, xmax= beta + se), size= 0.1, alpha= 0.7, color= colorBlindBlack8[2]) +
geom_errorbar(aes(ymin= BETA - SE, ymax= BETA + SE), size= 0.1, alpha= 0.7, color= colorBlindBlack8[2]) +
geom_point(size= 2, shape=21, stroke= 0.1, alpha= 0.7, fill= colorBlindBlack8[2]) +
xlab('Effect on testosterone (women), nmol/L') +
ylab('Effect on gestational duration, days') +
theme_cowplot(font_size= 8) +
geom_abline(intercept= 0, slope= filter(mr, method== 'IVW', trait== 'Testosterone_fem_cluster') %>% pull(estimate), colour= '#d9d9d9') +
geom_abline(intercept= (filter(mr, method== '(intercept)', trait== 'Testosterone_fem_cluster') %>% pull(estimate))[1], slope= filter(mr, method== 'MR-Egger', trait== 'Testosterone_fem_cluster') %>% pull(estimate), colour= '#d9d9d9', linetype= 'dashed') +
geom_hline(yintercept= 0, size= 0.1) +
geom_vline(xintercept= 0, size= 0.1) +
theme(axis.line.x = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks= element_blank(),
        panel.grid.major= element_line(colour= 'grey', size= 0.05))

ggsave(snakemake@output[[1]], plot= p1, width= 70, height= 70, units= 'mm', dpi= 300)
ggsave(snakemake@output[[2]], plot= p2, width= 70, height= 70, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/MR_sex_hormones_GA.R 
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[1]])

nr= nrow(d)

d= filter(d, Enrichment_p< 0.05 / (nrow(d)- 1))

d$description= with(d, ifelse(Category== 'H3K27ac_HniszL2_0', 'H3K27ac', 
			ifelse(Category== 'SuperEnhancer_HniszL2_0', 'SuperEnhancer',
			ifelse(Category== 'Backgrd_Selection_StatL2_0', 'Background selection',
			ifelse(Category== 'CpG_Content_50kbL2_0', 'CpG content',
			ifelse(Category== 'BLUEPRINT_DNA_methylation_MaxCPPL2_0', 'DNA Methylation', NA))))))

d= arrange(d, desc(Enrichment_p))

d$description= factor(d$description, levels= unique(d$description))

p1= ggplot(data=d, aes(x= description, y= -log10(Enrichment_p))) +
geom_col(fill=colorBlindBlack8[2], alpha= 0.6) +
theme_cowplot(font_size= 10) +
ylab('Enrichment -log10(pvalue)') +
theme(axis.title.y=element_blank()) +
geom_hline(yintercept= -log10(0.05/ (nr -1)), linetype= 'dashed', colour= 'grey') +
coord_flip()

p2= ggplot(data=d, aes(x= description, y= Enrichment)) +
geom_col(fill=colorBlindBlack8[4], alpha= 0.6) +
theme_cowplot(font_size= 10) +
ylab('Enrichment (h2 / proportion of SNPs)') +
theme(axis.title.y=element_blank(),
        axis.text.y=element_blank(),
        axis.ticks.y=element_blank()) +
geom_hline(yintercept= 1, linetype= 'dashed', colour= 'grey') +
coord_flip()

x= plot_grid(p1, p2)


ggsave(snakemake@output[[1]], plot= x, height= 50, width= 140, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 2 of figures/partitioned_h2.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)




d= fread(snakemake@input[[1]], h= T, select= c('ID', 'pvalue', 'EAF'))
d$MAF= ifelse(d$EAF>0.5,  1 - d$EAF, d$EAF)
d= arrange(d, pvalue)
d= d[!duplicated(d$ID), ]


d= mutate(d, maf_tertiles = ntile(MAF, 3))
#m1= round(max(d[d$maf_tertiles== 1, 'MAF']), 3)
#m2= round(max(d[d$maf_tertiles== 2, 'MAF']), 3)


#d$maf_tertiles= factor(d$maf_tertiles, levels=c("1", "2", "3"), labels=c(paste('MAF<', m1), paste(m1,'< MAF >', m2), paste('MAF>', m2)))

df= arrange(d, pvalue) %>% mutate(exp1= -log10(1:length(pvalue)/length(pvalue)))

p1= ggplot(filter(df, pvalue<0.05), aes(exp1, -log10(pvalue))) +
  geom_point(size= 0.4, color= colorBlindBlack8[2]) +
#scale_color_manual(values= colorBlindBlack8[c(2,4,8)])+
  geom_abline(intercept = 0, slope = 1, alpha = .5) +
labs(colour="") +
theme_cowplot(font_size= 12) +
xlab('Expected (-log10(p-value))') +
ylab('Observed (-log10(p-value))') +
theme(legend.position= 'bottom') 
#guides(colour = guide_legend(override.aes = list(size=3)))

ggsave(snakemake@output[[1]], plot= p1, width= 120, height= 120, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/QQ_plot.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
library(data.table)
library(dplyr)
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library('showtext')


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

d= fread(snakemake@input[[1]])


d= filter(d, !grepl('BW', trait), !grepl('GA_fetal', trait), !grepl('male', trait))

d$trait= with(d, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis'))))))))))))))))))))))

pregnancy= c('Miscarriage', 'Pre-eclampsia')
uterus= c('Leiomyoma uterus', 'Pelvic Organ Prolapse', 'Endometriosis', 'Polycystic ovary syndrome')
fitness= c('Age at first birth', 'Number of live births')
hormonal= c('Age at menarche', 'Age at menopause', 'Testosterone (women)', 'SHBG (women)', 'CBAT (women)', 'Oestradiol (women)')

d$cluster= with(d, ifelse(trait %in% pregnancy, 'Pregnancy', ifelse(trait %in% uterus, 'Reproductive organs', ifelse(trait %in% fitness, 'Fitness', 'Sex-hormone related'))))

d$colour= with(d, ifelse(cluster== 'Pregnancy', colorBlindBlack8[3], ifelse(cluster== 'Reproductive organs', colorBlindBlack8[5], ifelse(cluster== 'Fitness', colorBlindBlack8[7], colorBlindBlack8[8]))))

d$GENE= apply(d[, 'locus'], 1, function(x) unlist(strsplit(x, '_'))[2])

d$GENE= with(d, ifelse(GENE== 'CDC42', 'CDC42/ WNT4', ifelse(GENE== 'HIVEP3', 'HIVEP3/ EDN2', ifelse(GENE== 'TET3', 'TET3/ DGUOK-AS1', ifelse(GENE== 'TCEA2', 'TCEA2/ OPRL1', GENE)))))

d$sig= ifelse(d$PP.H4.abf>0.5, '*', '')

d= arrange(d, cluster)

d$trait= factor(d$trait, levels= unique(d$trait))
traits= unique(d$trait)
colors <- filter(d, !duplicated(trait)) %>% arrange(trait) %>% pull(colour)

d$PP= ifelse(d$PP.H4.abf> d$PP.H3.abf, d$PP.H4.abf, -d$PP.H3.abf - d$PP.H4.abf)
d$PP2= ifelse(d$PP.H4.abf> d$PP.H3.abf, d$PP.H4.abf, d$PP.H3.abf)
p1= ggplot(d, aes(trait, GENE, value= PP, fill= PP, colour= PP, size= PP2, stroke= 1-  PP)) + 
theme_cowplot(font_size= 9) +
geom_point(shape= 15) + 
scale_fill_gradient2(low= colorBlindBlack8[4], mid= 'white', high= colorBlindBlack8[2], guide= F) +
scale_colour_gradient2(low= colorBlindBlack8[4], mid= 'white', high= colorBlindBlack8[2], guide= F) +
scale_size_continuous(range= c(1, 2.5), guide= F) +
scale_x_discrete(position= 'top') +
theme(axis.ticks= element_blank(),
	axis.title= element_blank(),
	axis.text.x= element_blank())  +
geom_vline(xintercept= 1:(length(unique(d$trait))-1) + 0.5, size= 0.4, colour= 'grey') +
geom_hline(yintercept= 1:(length(unique(d$GENE))-1) + 0.5, size= 0.4, colour= 'grey') +
geom_vline(xintercept= cumsum(c(length(fitness) , length(pregnancy) , length(uterus) )) +0.5, size= 0.8) +
theme(	panel.grid.major = element_blank(), 
	panel.grid.minor = element_blank(),
	panel.background = element_blank(),
	panel.border = element_rect(colour= 'black', fill= NA, size=1),
	plot.margin = unit(c(0, 0.1, 0.1, 0), "cm"),
	axis.line= element_blank())

t_count_locus= group_by(d, trait) %>% summarize(PP= sum(as.numeric(PP.H4.abf> 0.8)), PP_locus= sum(as.numeric(PP.H4.abf + PP.H3.abf>0.8)))
t_count_locus$PP= t_count_locus$PP_locus - t_count_locus$PP
t_count_locus$supp= 'Locus-level'

t_count= group_by(d, trait) %>% summarize(PP= sum(as.numeric(PP.H4.abf> 0.8)))
t_count$supp= 'Coloc'

t_count= bind_rows(t_count, t_count_locus)

t_count$trait= factor(t_count$trait, levels= unique(d$trait))
t_count$supp= factor(t_count$supp, levels= c('Locus-level','Coloc'))

p2= ggplot(t_count, aes(trait, -PP, fill= supp)) +
theme_cowplot(font_size= 8) +
geom_col(alpha= 0.7) +
geom_hline(yintercept= 0) +
scale_fill_manual(values= c(colorBlindBlack8[4], colorBlindBlack8[2]), guide= F) +
theme(	axis.line= element_blank(),
	panel.grid.major = element_blank(), 
	panel.grid.minor = element_blank(),
	panel.background = element_blank(),
	panel.border = element_rect(colour= 'black', fill= NA, size=1),
	axis.text.x= element_blank(),
	axis.ticks.x= element_blank(),
	axis.title= element_blank(),
	plot.margin = unit(c(0, 0, 0, 0.1), "cm")) +
scale_y_continuous(limits= c(-10, 0), expand= c(0,0), labels= seq(0, 10, 2), breaks= seq(0, -10, -2)) +
geom_vline(xintercept= cumsum(c(length(fitness) , length(pregnancy) , length(uterus) )) +0.5, size= 0.8)  +
geom_hline(yintercept= c(-4, -8), size= 0.3, linetype= 'dashed', colour= 'grey')

l_count_locus= group_by(d, GENE) %>% summarize(PP= sum(as.numeric(PP.H4.abf> 0.8)), PP_locus= sum(as.numeric(PP.H4.abf + PP.H3.abf>0.8)))
l_count_locus$PP= l_count_locus$PP_locus - l_count_locus$PP
l_count_locus$supp= 'Locus-level'

l_count= group_by(d, GENE) %>% summarize(PP= sum(as.numeric(PP.H4.abf> 0.8)))
l_count$supp= 'Coloc'

l_count= bind_rows(l_count, l_count_locus)

l_count$trait= factor(l_count$GENE, levels= unique(d$GENE))
l_count$supp= factor(l_count$supp, levels= c('Locus-level','Coloc'))

print('done')
p3= ggplot(l_count, aes(PP, GENE, fill= supp)) +
theme_cowplot(font_size= 8) +
geom_col(alpha= 0.7) +
geom_hline(yintercept= 0) +
scale_fill_manual(values= c(colorBlindBlack8[4], colorBlindBlack8[2]), guide= F) +
theme(	axis.line= element_blank(),
	panel.grid.major = element_blank(), 
	panel.grid.minor = element_blank(),
	panel.background = element_blank(),
	panel.border = element_rect(colour= 'black', fill= NA, size=1),
	axis.text.y= element_blank(),
	axis.ticks.y= element_blank(),
	axis.title= element_blank(),
	plot.margin = unit(c(0, 0.1, 0, 0), "cm")) +
scale_x_continuous(limits= c(0, 10), expand= c(0,0), labels= seq(0,10, 2), breaks= seq(0, 10, 2))

x1= plot_grid(p1, p3, nrow= 1, align= 'h', rel_widths= c(2, 0.5))
x2= plot_grid(p1, p2, nrow= 2, align= 'v', rel_heights= c(2, 0.3))

ggsave(snakemake@output[[1]], plot= x1, width= 127 - 1, height= 127 - 25 - 1, units= 'mm', dpi= 300)
ggsave(snakemake@output[[2]], plot= x2, width= 103 - 1, height= 127 - 25 - 1, units= 'mm', dpi= 300)

################## Genetic correlations

d= fread(snakemake@input[[2]])

d= filter(d, grepl('GAraw', p1), !grepl('BW', p2), !grepl('male', p2))
#d$p1= 'Gestational duration (maternal)'
d$p1= 'Maternal'
x= fread(snakemake@input[[2]])

x= filter(x, grepl('GA_fetal', p1), !grepl('BW', p2), !grepl('male', p2))
#x$p1= 'Gestational duration (fetal)'
x$p1= 'Fetal'
d= rbind(d, x)

d$p2= gsub('.txt.sumstats.gz', '', apply(d[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))
d$trait= d$p2

d$trait= with(d, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis'))))))))))))))))))))))


d= filter(d, trait!= 'GA fetal effect')

d$cluster= with(d, ifelse(trait %in% pregnancy, 'Pregnancy', ifelse(trait %in% uterus, 'Reproductive organs', ifelse(trait %in% fitness, 'Fitness', 'Sex-hormone related'))))

d$colour= with(d, ifelse(cluster== 'Pregnancy', colorBlindBlack8[3], ifelse(cluster== 'Reproductive organs', colorBlindBlack8[1], ifelse(cluster== 'Fitness', colorBlindBlack8[7], colorBlindBlack8[8]))))

d= arrange(d, cluster)

d$trait= factor(d$trait, levels= traits)

colors <- filter(d, !duplicated(trait)) %>% arrange(trait) %>% pull(colour)


d$sig= ifelse(d$p< 0.05/ (nrow(d)/2), '**', ifelse(d$p< 0.05, '*', ''))
d= filter(d, p1== 'Maternal')
d$p1= 'Gestational duration'

rg_plot= ggplot(d, aes(trait, p1, fill= rg)) +
geom_tile(colour = "white", size= 1) +
theme_cowplot(font_size= 9) +
scale_fill_gradient2(low= colorBlindBlack8[2], high= colorBlindBlack8[4], mid= 'white', guide= F) +
theme(axis.text.x = element_text(angle = 45, hjust = 0),
        axis.title.x = element_blank(),
        axis.title.y = element_blank()) +
scale_x_discrete(position = "top") +
geom_text(data= d, aes(trait, p1, label= sig), size= 6/ .pt) +
theme(  panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        panel.background = element_blank(),
	axis.ticks= element_blank(),
        panel.border = element_rect(colour= 'black', fill= NA, size=1),
        plot.margin = unit(c(0, 1, 0, 0), "cm"),
        axis.line= element_blank(),
	axis.text.x= element_text(angle= 45, hjust=0, colour= colors))


x2= plot_grid(rg_plot,p1, nrow= 2, align= 'v', rel_heights= c(0.85, 2))

ggsave(snakemake@output[[3]], plot= x2, width= 113 - 2.5, height= 127 - 25 - 1 , units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/repr_pheno_coloc.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(tidyverse)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")


font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

x= fread(snakemake@input[[1]])

x$p1= gsub('.txt.sumstats.gz', '', apply(x[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x$p2= gsub('.txt.sumstats.gz', '', apply(x[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))

x1= fread(snakemake@input[[2]])

x1$p1= gsub('.txt.sumstats.gz', '', apply(x1[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x1$p2= gsub('.txt.sumstats.gz', '', apply(x1[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))
x1$rg= -1 * x1$rg
d= rbind(x, x1)

#traits= filter(d, p< 0.05/ 14, !grepl('BW', p2), !grepl('GA', p2)) %>% pull(p2)

d$trait= d$p2
d$trait= with(d, ifelse(trait== 'GAraw', 'Maternal gestational duration',
ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal BW',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Estradiol (women)',
                ifelse(trait== 'POP', 'Pelvic organ prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis')))))))))))))))))))))))

d= filter(d, !grepl('BW', p2), !grepl('GA', p2), !grepl('_male', p2))

traits= unique(arrange(d, p) %>% pull(trait))
d$trait= factor(d$trait, levels= rev(traits))

p1= ggplot(d, aes(rg, trait, colour= p1)) + 
geom_pointrange(aes(xmax= rg + 1.96 * se, xmin= rg - 1.96 * se), position = position_dodge(width = 0.3), fatten= 1) +
scale_colour_manual(values= colorBlindBlack8[c(8,3)], guide= FALSE) +
theme_cowplot(font_size= 9) +
scale_x_continuous(limits= c(-1, 1), breaks= seq(-1, 1, 0.5)) +
xlab('Genetic correlation') +
geom_vline(xintercept= 0, size= 0.3) +
geom_vline(xintercept= c(seq(-1, 1, 0.25)), colour= 'grey', linetype= 'dashed', alpha= 0.5, size= 0.2) +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3),
        axis.title.y= element_blank())


ggsave(snakemake@output[[1]], plot= p1, width= 88, height= 120, units= 'mm', dpi= 300)

fwrite(d, snakemake@output[[2]], sep= '\t')

p1= ggplot(d, aes(rg, trait, colour= p1)) +
geom_pointrange(aes(xmax= rg + 1.96 * se, xmin= rg - 1.96 * se), position = position_dodge(width = 0.3), fatten= 1) +
scale_colour_manual(values= colorBlindBlack8[c(8,3)], name= 'Trait') +
theme_cowplot(font_size= 9) +
scale_x_continuous(limits= c(-1, 1), breaks= seq(-1, 1, 0.5)) +
xlab('Genetic correlation') +
geom_vline(xintercept= 0, size= 0.3) +
geom_vline(xintercept= c(seq(-1, 1, 0.25)), colour= 'grey', linetype= 'dashed', alpha= 0.5, size= 0.2) +
theme(axis.line.x = element_line(size = 0.3),
        axis.line.y = element_line(size = 0.3),
        axis.ticks= element_line(size= 0.3),
        axis.title.y= element_blank())


ggsave(snakemake@output[[3]], plot= p1, width= 88, height= 120, units= 'mm', dpi= 300)
R ggplot2 tidyverse dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/repr_pheno_correlations.R 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
library(scales)
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
library(tidyverse)
library(fmsb)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")


font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)


x= fread(snakemake@input[[1]])
x$p1= gsub('.txt.sumstats.gz', '', apply(x[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x$p2= gsub('.txt.sumstats.gz', '', apply(x[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))

x1= fread(snakemake@input[[2]])
x1$p1= gsub('.txt.sumstats.gz', '', apply(x1[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x1$p2= gsub('.txt.sumstats.gz', '', apply(x1[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))

x= rbind(x, x1)

traits= unique(filter(x, p< 0.05/ 14, !grepl('BW', p2), !grepl('GA', p2)) %>% pull(p2))

d= fread(snakemake@input[[3]])

table_supp= d
table_supp$pheno= 'Gestational duration'
d$gcp.pm= ifelse(d$pval.gcpzero.2tailed< 0.05/length(traits), d$gcp.pm, 0)

d= filter(d, repr_pheno %in% traits)

d= arrange(d, desc(gcp.pm))

df= fread(snakemake@input[[4]])

table_supp2= df
table_supp2$pheno= 'Preterm delivery' 

table_supp= rbind(table_supp, table_supp2)

df$gcp.pm= ifelse(df$pval.gcpzero.2tailed< 0.05/length(traits), df$gcp.pm, 0)

df= filter(df, repr_pheno %in% traits)

d= inner_join(d, df, by= 'repr_pheno')
d$trait= d$repr_pheno
d$trait= with(d, ifelse(trait== 'GAraw', 'Maternal gestational duration',
ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal BW',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis')))))))))))))))))))))))

d$repr_pheno= d$trait
x= as.data.frame(matrix(d$gcp.pm.x, ncol= nrow(d)))
x=rbind(x, as.data.frame(matrix(d$gcp.pm.y, ncol= nrow(d))))



names(x)= d$repr_pheno
rownames(x)= c('Preterm delivery', 'Gestational duration ')
x= rbind(rep(1,nrow(d)) , rep(0,nrow(d)) , x)

inches= 25.4

pdf(snakemake@output[[1]], width= 88 / inches, height= 88 / inches)
par(mar=c(0,0,0,0))


radarchart(abs(x), axistype= 0,

    #custom polygon
    pcol= c(colorBlindBlack8[3], colorBlindBlack8[8]) , pfcol= c(alpha(colorBlindBlack8[3], 0.4), alpha(colorBlindBlack8[8], 0.4)) , plwd=1, pty= 16, plty= 1, vlcex= 0.8, vlabels= c('Testosterone\n(women)', 'Age at\nfirst birth', 'Age at\nmenopause', 'Number of\nlive births', 'SHBG\n(women)', 'CBAT\n(women)'),
    #custom the grid
    cglcol="grey", cglty=1, axislabcol="#525252", caxislabels= seq(0, 1, 0.25), cglwd=0.8, calcex= 0.4

    #custom labels
    )

dev.off()

table_supp$trait= table_supp$repr_pheno
table_supp$trait= with(table_supp, ifelse(trait== 'GAraw', 'Maternal gestational duration',
ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal BW',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycystic ovary syndrome', 'Endometriosis')))))))))))))))))))))))


fwrite(table_supp, snakemake@output[[2]], sep= '\t')
R tidyverse dplyr data.table tidyr cowplot ggrepel knitr scales showtext From line 1 of figures/repr_pheno_LCV.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')
options(warn=-1)


d= fread(snakemake@input[[1]], h= T)


colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

font_add("arial", "arial.ttf", bold= 'arial_bold.ttf')

showtext_opts(dpi = 300)
showtext_auto(enable = TRUE)

female_repr= c('breast', "cervix, uterine", 'endometrium', 'ovary', 'placenta', 'vagina', 'fallopian tube')
male_repr= c('ductus deferens', 'testis', 'seminal vesicle', 'prostate', 'epididymis')
muscle= c('smooth muscle', 'heart muscle', 'skeletal muscle')

d$organ= with(d, ifelse(tissue %in% female_repr, 'Female reproductive', ifelse(tissue %in% male_repr, 'Male reproductive', ifelse(tissue %in% muscle, 'Muscle', 'Others'))))

p1= ggplot(d, aes(-log10(MannW_pvalue), I(i_listmedian/ base_list_median), colour= organ)) +
geom_point(size= 1.5) +
theme_cowplot(font_size= 8) +
scale_colour_manual('Legend', values= c(colorBlindBlack8[c(3, 2, 8)], 'grey'), guide= 'none') +
geom_text_repel(data= filter(d, MannW_pvalue< 0.05), aes(label= tissue), fontface = 'bold') +
geom_vline(xintercept= -log10(0.05), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_vline(xintercept= -log10(0.05/nrow(d)), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
ylab('Enrichment') +
xlab('-log10(pvalue)')


ggsave(snakemake@output[[1]], plot= p1, width= 120, height= 90, units= 'mm', dpi= 300)


p1= ggplot(d, aes(-log10(MannW_pvalue), I(i_listmedian/ base_list_median), colour= organ)) +
geom_point(size= 1.5) +
theme_cowplot(font_size= 10) +
scale_colour_manual('Legend', values= c(colorBlindBlack8[c(3, 2, 8)], 'grey')) +
geom_text_repel(data= filter(d, MannW_pvalue< 0.05), aes(label= tissue), fontface = 'bold', show_guide = FALSE) +
geom_vline(xintercept= -log10(0.05), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
geom_vline(xintercept= -log10(0.05/nrow(d)), colour= colorBlindBlack8[8], linetype= 'dashed', size= 0.2, alpha= 0.6) +
ylab('Enrichment') +
xlab('-log10(pvalue)')

ggsave(snakemake@output[[2]], plot= p1, width= 120, height= 90, units= 'mm', dpi= 300)
R ggplot2 dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of figures/RNA_enrichment.R 
11
12
script:
	'manhattan_plot.R'
SnakeMake From line 11 of figures/Snakefile 
25
26
script:
	'lm_effect_origin.R'
SnakeMake From line 25 of figures/Snakefile 
37
38
script:
	'effect_origin_dendrogram.R'
SnakeMake From line 37 of figures/Snakefile 
48
49
script:
	'effect_origin_ternary.R'
SnakeMake From line 48 of figures/Snakefile 
64
65
script:
	'gene_based_vs_coloc_iPSC.R'
SnakeMake From line 64 of figures/Snakefile 
74
75
script:
	'BW_coloc_spider.R'
SnakeMake From line 74 of figures/Snakefile 
87
88
script:
        'KCNAB1_pheWAS.R'
SnakeMake From line 87 of figures/Snakefile 
101
102
script:
	'ADCY5_pheWAS.R'
SnakeMake From line 101 of figures/Snakefile 
118
119
script:
	'ADCY5_FST_AFR_EUR.R'
SnakeMake From line 118 of figures/Snakefile 
129
130
script:
	'BW_genetic_correlations.R'
SnakeMake From line 129 of figures/Snakefile 
143
144
script:
        'repr_pheno_correlations.R'
SnakeMake From line 143 of figures/Snakefile 
153
154
script:
	'partitioned_h2.R'
SnakeMake From line 153 of figures/Snakefile 
164
165
script:
	'MacArthurlab_enrichment.R'
SnakeMake From line 164 of figures/Snakefile 
180
181
script:
	'ADCY5_effect_direction.R'
SnakeMake From line 180 of figures/Snakefile 
193
194
script:
	'BW_conditioning.R'
SnakeMake From line 193 of figures/Snakefile 
207
208
script:
        'BW_conditioning_top.R'
SnakeMake From line 207 of figures/Snakefile 
218
219
script:
        'mediation_BW_GA_individual_level_data.R'
SnakeMake From line 218 of figures/Snakefile 
230
231
script:
        'mediation_BW_GA_individual_level_data_decode.R'
SnakeMake From line 230 of figures/Snakefile 
245
246
script:
	'MR_GA_BW_haplotype.R'
SnakeMake From line 245 of figures/Snakefile 
257
258
script:
	'repr_pheno_coloc.R'
SnakeMake From line 257 of figures/Snakefile 
270
271
script:
	'repr_pheno_LCV.R'
SnakeMake From line 270 of figures/Snakefile 
282
283
script:
	'repr_pheno_correlations.R'
SnakeMake From line 282 of figures/Snakefile 
292
293
script:
	'RNA_enrichment.R'
SnakeMake From line 292 of figures/Snakefile 
305
306
script:
	'QQ_plot.R'
SnakeMake From line 305 of figures/Snakefile 
316
317
script:
	'h2_allphenos.R'
SnakeMake From line 316 of figures/Snakefile 
326
327
script:
	'h2_cohorts.R'
SnakeMake From line 326 of figures/Snakefile 
335
336
script:
        'genet_correlations_meta.R'
SnakeMake From line 335 of figures/Snakefile 
346
347
script:
	'manhattan_plot_postTerm.R'
SnakeMake From line 346 of figures/Snakefile 
357
358
script:
        'manhattan_plot_postTerm.R'
SnakeMake From line 357 of figures/Snakefile 
368
369
script:
	'forest_plot_EEFSEC.R'
SnakeMake From line 368 of figures/Snakefile 
380
381
script:
	'MR_sex_hormones_GA.R'
SnakeMake From line 380 of figures/Snakefile 
391
392
script:
	'cell_type_enrichment.R'
SnakeMake From line 391 of figures/Snakefile 
402
403
script:
	'labor_deg.R'
SnakeMake From line 402 of figures/Snakefile 
413
414
script:
	'coloc_sex_hormones.R'
SnakeMake From line 413 of figures/Snakefile 
423
424
script:
	'evo.R'
SnakeMake From line 423 of figures/Snakefile 
432
433
script:
	'GA_BW_PGS_correlations.R'
SnakeMake From line 432 of figures/Snakefile 
445
446
script:
	'GAraw_vs_allPTD.R'
SnakeMake From line 445 of figures/Snakefile 
11
12
13
14
15
run:
	d= pd.read_csv(input[0], sep= '\t', header= None, names= ['CHR', 'start', 'end', 'geneSymbol', 'Ensembl_gene'])
	d= d.loc[~d.geneSymbol.str.contains(' '), :]
	d= d[['CHR', 'start', 'end', 'geneSymbol']]
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 11 of gene_based/Snakefile 
24
25
26
27
28
29
run:
	d= pd.read_csv(input[0], sep='\t', header= 0)
	pop= ['CEU', 'TSI', 'FIN', 'GBR', 'IBS']
	d= d.loc[d.Population.isin(pop)]
	d['IID']= d['Individual ID']
	d.to_csv(output[0], sep= '\t', header= False, index= False, columns= ['IID'])
SnakeMake From line 24 of gene_based/Snakefile 
37
38
39
run:
	vcfs= [x for x in input if '1000g' in x]
	shell('/home/pol/software/bcftools-1.9/bcftools concat {input} -o {output[0]} -Oz')
SnakeMake BCFtools From line 37 of gene_based/Snakefile 
52
53
shell:
	'/home/pol/software/plink2 --vcf {input[0]} --max-alleles 2 --keep {input[1]} --make-bed --out {params[0]}'
SnakeMake plink2 From line 52 of gene_based/Snakefile 
65
66
67
68
69
70
71
72
73
74
75
run:
	d= pd.read_csv(input[0], sep= '\t', header= None, names= ['CHR', 'RSID', 'cm', 'POS', 'A1', 'A2'])
	d['REF']= np.where(d.A1.str.len() > d.A2.str.len(), 'I', d.A1)
	d['EFF']= np.where(d.A2.str.len() > d.A1.str.len(), 'I', d.A2)
	d['REF']= np.where(d.EFF== 'I', 'D', d.REF)
	d['EFF']= np.where(d.REF== 'I', 'D', d.EFF)
	d['RSID']= np.where(d.REF > d.EFF, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.EFF + ':' + d.REF, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.REF + ':' + d.EFF)
	d= d[['CHR', 'RSID', 'cm', 'POS', 'A1', 'A2']]
	d.to_csv(output[0], sep= '\t', header= False, index= False)
	shell('mv {input[1]} {output[1]}')
	shell('mv {input[2]} {output[2]}')
SnakeMake From line 65 of gene_based/Snakefile 
83
84
85
86
87
88
89
90
run:
	d= pd.read_csv(input[0], header= 0, sep= '\t', compression= 'gzip', usecols= ['ID', 'pvalue'])
	d.dropna(subset= ['ID'], inplace= True)
	d= d.loc[d.ID != '-', :]
	d= d[['ID', 'pvalue']]
	d.columns= ['SNP', 'p']
	d['SNP']= d.SNP.str.replace('^23:', 'X:')
	d.to_csv(output[0], sep= '\t', header= True, index= None, columns= ['SNP', 'p'])
SnakeMake From line 83 of gene_based/Snakefile 
105
106
shell:
	'/home/pol/software/gcta_1.93.2beta/gcta64 --bfile {params[0]} --maf 0.01 --fastBAT {input[1]} --fastBAT-gene-list {input[2]} --out {params[1]} --thread-num {threads}'
SnakeMake From line 105 of gene_based/Snakefile 
12
13
14
15
run:
	d=pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'RSID', 'pvalue'])[['RSID', 'CHR', 'POS', 'pvalue']]
	d.columns= ['SNP', 'CHR', 'POS', 'P']
	d.to_csv(output[0], header= True, index= False, sep= '\t')
SnakeMake From line 12 of independent/Snakefile 
23
24
25
26
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['CHR', 'POS'])
        d.sort_values(['CHR', 'POS'], inplace= True)
        d['pos2']= d.POS
SnakeMake From line 23 of independent/Snakefile 
38
39
40
41
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d= d.loc[d.Relationship== 'unrel', :]
	pop= ['CEU', 'TSI', 'FIN', 'GBR', 'IBS']
SnakeMake From line 38 of independent/Snakefile 
54
55
56
run:
	vcfs= [infile for infile in input if 'vcf' in infile]
	shell('/home/pol/software/bcftools-1.9/bcftools concat -a -O v -R {input[0]} {vcfs} -o {output[0]}')
SnakeMake BCFtools From line 54 of independent/Snakefile 
68
69
shell:
	'/home/pol/software/plink --vcf {input[0]} --keep {input[1]} --make-bed -out {params[0]}'
SnakeMake pLink From line 68 of independent/Snakefile 
77
78
79
80
81
run:
	d= pd.read_csv(input[0], sep= '\t', header= None, names= ['chr', 'snp', 'x1', 'pos', 'a1', 'a2'])
	d= d[d.duplicated(['snp'], keep=False)]
	d.drop_duplicates(subset= ['snp'], keep= 'first')
	d.to_csv(output[0], sep= '\t', header= False, index= False)
SnakeMake From line 77 of independent/Snakefile 
94
95
shell:
	'~/software/plink --bfile {params[0]} --clump {input[0]} --exclude {input[1]} --clump-r2 0.05 --clump-kb 1000 --clump-p1 5e-8 --clump-p2 1e-5 --out {params[1]}'
SnakeMake pLink From line 94 of independent/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

df= fread(snakemake@input[[1]], select= c('ID', 'BETA', 'SE', 'TOTALSAMPLESIZE', 'EAF'))
df$MAF= ifelse(df$EAF>0.5, 1 - df$EAF, df$EAF)

z= fread(snakemake@input[[2]])
z$n= 716
df= inner_join(df, z, by= 'ID')

rm(z)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]

if (!grepl('sQTL', snakemake@output[[1]])) {

cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tprotein\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tprotein\n', file= snakemake@output[[2]])
} else {

cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tprotein\tgene\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tprotein\tgene\n', file= snakemake@output[[2]])

}

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6


df= data.frame(df)

colocalization_eqtl= function(temp_df){
	protein= unique(temp_df$gene)
        if (nrow(temp_df)== 0) {

        PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')


        } else {
	temp_df = filter(temp_df, SE>0, se> 0)

	if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID,s= 0.067)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, s= 0.122)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID) }

        data2= list(beta= temp_df$beta, varbeta= temp_df$se**2, N=temp_df$n, type= 'quant', snp= temp_df$ID)
        myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, MAF=temp_df$MAF, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
        if (length(myres)==1 ) { 
        PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
        } else {
        PPH= data.frame(t(myres[[1]]))
        PPH$protein= protein
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$protein= protein
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
}
}
}



mclapply(split(df, df$gene), colocalization_eqtl, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of iPSC/coloc_iPSC.R 
15
16
script:
	'coloc_iPSC.R'
SnakeMake From line 15 of iPSC/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
library(dplyr)
library(data.table)

d= fread(snakemake@input[[1]])
d= filter(d, !is.na(Z))

x= fread(snakemake@input[[2]])
x= filter(x, !is.na(Z))

ld= fread(snakemake@input[[3]])

d= inner_join(d, x, by= 'SNP')
d= inner_join(d, ld, by= 'SNP')


source(snakemake@params[[1]])
setwd(snakemake@params[[2]])

LCV= RunLCV(d$L2, d$Z.y, d$Z.x, ldsc.intercept= 1, n.1= (d$N.y), n.2= (d$N.x))

cat('zscore\tpval.gcpzero.2tailed\tgcp.pm\tgcp.pse\trho.est\trho.err\tpval.fullycausal1\tpval.fullycausal2\th2.zscore1\th2.zscore2\tpheno\trepr_pheno\n', file = snakemake@output[[1]])

z= data.frame(zscore= LCV$zscore, pval.gcpzero.2tailed= LCV$pval.gcpzero.2tailed, gcp.pm= LCV$gcp.pm, gcp.pse= LCV$gcp.pse, rho.est= LCV$rho.est, rho.err= LCV$rho.err, pval.fullycausal1= LCV$pval.fullycausal[1],pval.fullycausal2= LCV$pval.fullycausal[2], h2.zscore1= LCV$h2.zscore[1], h2.zscore2= LCV$h2.zscore[2], pheno= snakemake@wildcards[['pheno']], repr_pheno= snakemake@wildcards[['repr_pheno']])

fwrite(z, snakemake@output[[1]], sep= '\t')
R dplyr data.table From line 1 of LCV/LCV.R 
13
14
script:
	'LCV.R'
SnakeMake From line 13 of LCV/Snakefile 
22
23
24
25
26
	shell:
		'''
		head -1 {input[0]} > {output[0]}
                tail -n +2 -q {input} >> {output[0]}
		'''
SnakeMake From line 22 of LCV/Snakefile 
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
import pandas as pd
import numpy as np
from scipy.special import chdtri
import gzip
import csv

def not_number(s):
	if s != None:
		try:
			float(s)
			return False
		except ValueError:
			return True
	else:
		return True


def select_format(repr_pheno, row):
	'For each wildcard assign the correct formating function.'
	if repr_pheno== 'Preeclampsia':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= preeclampsia(row)
	if repr_pheno== 'POP': 
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= POP(row)
	if repr_pheno== 'miscarriage':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= miscarriage(row)
	if repr_pheno== 'GA_fetal':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= fet_GA(row)
	if repr_pheno== 'BW_maternal':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= BW_maternal(row)
	if repr_pheno== 'BW_fetal':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= BW_fetal(row)
	if repr_pheno== 'BW_maternal_effect':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= BW_maternal_adjusted_effect(row)
	if repr_pheno== 'BW_fetal_effect':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= BW_fetal_adjusted_effect(row)
	if repr_pheno== 'leiomyoma_uterus':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= leiomyoma_uterus(row)
	if repr_pheno== 'AMenopause':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= AMenopause(row)
	if repr_pheno in ['Oestradiol_fem', 'NLB', 'AFB', 'AMenarche', 'endometriosis']:
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= UKBB_traits(row)
	if repr_pheno in ['SHBG_fem', 'Testosterone_fem', 'Testosterone_male', 'SHBG_male', 'CBAT_fem', 'CBAT_male']:
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= pritchard(row)
	if repr_pheno == 'PCOS':
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= PCOS(row)
	if repr_pheno in ['Ruth_CBAT_female', 'Ruth_CBAT_male', 'Ruth_SHBG_female', 'Ruth_SHBG_male', 'Ruth_Testosterone_female', 'Ruth_Testosterone_male', 'Ruth_oestradiol']:
		rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= Ruth(row, repr_pheno) 
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]


def AMenopause(row):
	'REPROGEN Age at menopause.'
	EAF= float(row['EAF'])
	CHR= row['CHR']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['POS'])
	REF= row['Other_Allele'].upper()
	EFF= row['Effect_Allele'].upper()
	BETA= float(row['Effect'])
	pvalue= float(row['Pval'])
	SE= float(row['SE'])
	N= int(row['N'])
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def Ruth(row, repr_pheno):
	''
	EAF= float(row['effect_allele_frequency'])
	CHR= row['chromosome']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['base_pair_location'])
	REF= row['other_allele']
	EFF= row['effect_allele']
	BETA= float(row['beta'])
	pvalue= float(row['p_value'])
	SE= float(row['standard_error'])
	N= np.where(repr_pheno== 'Ruth_SHBG_female', 189473,
	np.where(repr_pheno== 'Ruth_SHBG_make', 180726,
	np.where(repr_pheno== 'Ruth_Testosterone_female', 230454,
	np.where(repr_pheno== 'Ruth_SHBG_male',194453 ,
	np.where(repr_pheno== 'Ruth_CBAT_female', 188507,
	np.where(repr_pheno== 'Ruth_SHBG_male', 178782, 206927))))))
	rsid= row['variant_id']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def pritchard(row):
	''
	EAF= float(row['A1_FREQ'])
	CHR= row['#CHROM']
	if CHR== 'X': CHR= 23
	if not_number(CHR): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	POS= int(row['POS'])
	CHR= int(CHR)
	REF= row['REF']
	EFF= row['ALT']
	N= int(row['OBS_CT'])
	if not_number(row['BETA']): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	if not_number(row['SE']): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	if not_number(row['P']): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	BETA= float(row['BETA'])
	SE= float(row['SE'])
	pvalue= float(row['P'])
	rsid= row['ID']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def leiomyoma_uterus(row):
	''
	EAF= float(row['EAF'])
	CHR= row['CHR']
	if CHR== 'X': CHR= 23
	if not_number(CHR): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	POS= int(row['POS'])
	CHR= int(CHR)
	REF= row['REF']
	EFF= row['EFF']
	N= row['TOTALSAMPLESIZE']
	BETA= float(row['beta'])
	SE= float(row['se'])
	pvalue= float(row['pvalue'])
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def preeclampsia(row):
	''
	CHR= row['CHR']
	if CHR== 'X': CHR= 23
	if not_number(CHR): return [0, 0, 0 , 0, 0, 0, 0, 0, 0, 0]
	POS= int(row['POS'])
	CHR= int(CHR)
	REF= row['REF'].upper()
	EFF= row['EFF'].upper()
	N= 4630 + 373345
	rsid= row['rsid']
	BETA= float(row['beta'])
	SE= float(row['se'])
	EAF= float(row['EAF'])
	pvalue= float(row['pvalue'])
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def BW_fetal_adjusted_effect(row):
	'Define each header for Birth weight fetal effect.'
	EAF= float(row['eaf'])
	CHR= row['chr']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['pos'])
	REF= row['nea'].upper()
	if REF== 'R': REF= 'D'
	EFF= row['ea'].upper()
	if EFF== 'R': EFF= 'D'
	BETA= float(row['beta'])
	pvalue= float(row['p'])
	SE= float(row['se'])
	N= int(row['n_ownBW'])
	rsid= row['RSID']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def BW_maternal_adjusted_effect(row):
	'Define each header for Birth weight fetal effect.'
	EAF= float(row['eaf'])
	CHR= row['chr']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['pos'])
	REF= row['nea'].upper()
	if REF== 'R': REF= 'D'
	EFF= row['ea'].upper()
	if EFF== 'R': EFF= 'D'
	BETA= float(row['beta'])
	pvalue= float(row['p'])
	SE= float(row['se'])
	N= int(row['n_offBW'])
	rsid= row['RSID']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]


def BW_maternal(row):
	'Define each header for Birth weight maternal effect.'
	EAF= float(row['eaf'])
	CHR= row['chr']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['pos'])
	REF= row['nea']
	EFF= row['ea']
	if REF== 'R': REF= 'D'
	if EFF== 'R': EFF= 'D'
	BETA= float(row['beta'])
	pvalue= float(row['p'])
	SE= float(row['se'])
	N= int(row['n'])
	rsid= row['SNP']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def BW_fetal(row):
	'Define each header for Birth weight maternal effect.'
	EAF= float(row['eaf'])
	CHR= row['chr']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['pos'])
	REF= row['nea']
	EFF= row['ea']
	if REF== 'R': REF= 'D'
	if EFF== 'R': EFF= 'D'
	BETA= float(row['beta'])
	pvalue= float(row['p'])
	SE= float(row['se'])
	N= int(row['n'])
	rsid= row['rsid']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]


def PCOS(row):
	'Define each header for PCOS excluding 23andme.'
	EAF= float(row['EAF'])
	CHR= row['CHR']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['POS'])
	REF= row['REF']
	EFF= row['EFF']
	BETA= float(row['beta'])
	pvalue= float(row['pvalue'])
	SE= float(row['se'])
	N= int(round(float(row['TOTALSAMPLESIZE'])))
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def UKBB_traits(row):
	'Define each header for UKBB traits (hormones).'
	if row['low_confidence_variant']== 'true': return [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
	CHR= row['variant'].split(':')[0]
	if CHR== 'X': CHR= 23
	POS= row['variant'].split(':')[1]
	if any([not_number(t) for t in [row['minor_AF'], CHR, POS, row['beta'], row['pval'], row['se'], row['n_complete_samples']]]): return [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
	CHR= int(CHR)
	POS= int(POS)
	REF= row['variant'].split(':')[2]
	EFF= row['variant'].split(':')[3]
	BETA= float(row['beta'])
	pvalue= float(row['pval'])
	SE= float(row['se'])
	N= int(row['n_complete_samples'])
	if row['minor_allele']== EFF:
		EAF= float(row['minor_AF'])
	else:
		EAF= 1- float(row['minor_AF'])
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def AP_repr(row):
	'Define each header for BOLT-LMM sumstats.'
	EAF= float(row['EAF'])
	CHR= row['CHR']
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['POS'])
	REF= row['A2']
	EFF= row['A1']
	BETA= float(row['Beta'])
	pvalue= float(row['P'])
	SE= float(row['se'])
	N= row['N']
	rsid= row['SNP']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]


def POP(row):
	'Define each header for pelvic organ prolapse.'
	if not row['CHR'].isdigit(): return [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
	EAF= float(row['EAF'])
	MAF= np.where(EAF> 0.5, 1 - EAF, EAF)
	if MAF < 0.005: return [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
	if row['CHR']== 'X': row['CHR']= 23
	CHR= int(row['CHR'])
	POS= int(row['POS'])
	REF= row['REF']
	EFF= row['EFF']
	BETA= float(row['BETA'])
	pvalue= float(row['pvalue'])
	SE= float(row['SE'])
	N= float(row['N'])
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def fet_GA(row):
	'Define each header for Fetal gestational duration.'
	EAF= ''
	if row['Chr']== 'X': row['Chr']= 23
	CHR= int(row['Chr'])
	POS= int(row['Pos'])
	REF= row['Non_effect_allele'].upper()
	EFF= row['Effect_allele'].upper()
	BETA= float(row['Effect'])
	pvalue= float(row['P'])
	SE= float(row['StdErr'])
	N= int(row['N'])
	rsid= row['Rsid']
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]

def miscarriage(row):
	'Define each header for Miscarriage.'
	EAF= row['Freq1']
	CHR= row['MarkerName'].split(':')[0]
	if CHR== 'X': CHR= 23
	CHR= int(CHR)
	POS= int(row['MarkerName'].split(':')[1])
	REF= row['Allele2'].upper()
	EFF= row['Allele1'].upper()
	BETA= float(row['Effect'])
	pvalue= float(row['P-value'])
	SE= float(row['StdErr'])
	N= 49996 + 174109
	rsid= ''
	return [rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue]


def format_list(input, output):
	with gzip.open(input, 'rt', newline='') as f:
		print(input)
		dialect = csv.Sniffer().sniff(f.readline(), delimiters= ' \t')
		f.seek(0)
		input_file= csv.DictReader(f, dialect= dialect)
		df_list= list()
		with open(output, 'w') as csvfile:
			writer = csv.writer(csvfile, delimiter= '\t')
			writer.writerow([g for g in ['ID', 'rsid', 'CHR', 'POS', 'EAF', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']])
		for row in input_file:
			rsid, CHR, POS, EAF, N, REF, EFF, BETA, SE, pvalue= select_format(snakemake.wildcards.repr_pheno, row)
			if CHR== 0: continue
			if len(REF) >1: REF= 'I'
			if len(EFF) >1: EFF= 'I'
			if REF== 'I': EFF= 'D'
			if EFF== 'I': REF= 'D'
			if REF> EFF:
				ID= str(CHR) + ':' + str(POS) + ':' + EFF + ':' + REF
				BETA= -1 * float(BETA)
				ref= EFF
				eff= REF
				EAF= 1 - float(EAF)
			else:
				ID= str(CHR) + ':' + str(POS) + ':' + REF + ':' + EFF
				BETA= float(BETA)
				eff= EFF
				ref= REF
			df_list.append([ID, rsid, CHR, POS, EAF, N, ref, eff, BETA, SE, pvalue])
			if len(df_list)== 1000:
				with open(output, 'a', newline= '') as file_handler:
					writer1= csv.writer(file_handler, delimiter= '\t')
					for item in df_list:
						writer1.writerow(item)
				df_list= list()
	with open(output, 'a', newline= '') as file_handler:
			writer1= csv.writer(file_handler, delimiter= '\t')
			for item in df_list:
				writer1.writerow(item)


format_list(snakemake.input[0], snakemake.output[0])
Python Pandas numpy scipy preeclampsia From line 1 of LDscore_reprpheno/format_sumstats.py 
8
9
script:
	'format_sumstats.py'
SnakeMake From line 8 of LDscore_reprpheno/Snakefile 
21
22
23
24
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d.columns= ['ID', 'SNP', 'CHR', 'POS', 'EAF', 'N', 'A2', 'A1', 'BETA', 'SE', 'pvalue']
	d.dropna(subset= ['pvalue'], axis= 0, inplace= True)
SnakeMake From line 21 of LDscore_reprpheno/Snakefile 
53
54
55
56
57
58
59
60
61
62
63
64
65
        shell:
                """
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
		conda activate ldsc
                python2 /home/pol/software/ldsc/munge_sumstats.py \
                --out {params[0]} \
		--merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
                --sumstats {input[0]} \
                --chunksize 500000
                conda deactivate
                set -eu
                """
SnakeMake From line 53 of LDscore_reprpheno/Snakefile 
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
	run:
                allfiles= [infile for infile in input if wildcards.pheno not in infile]
                allfiles= ','.join(allfiles)
                outfile= params[0] + wildcards.pheno + '_rg'
                infile= input[0]
                shell("""
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --rg {infile},{allfiles} \
                --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --out {outfile}
                conda deactivate
                set -eu
                """)
SnakeMake From line 77 of LDscore_reprpheno/Snakefile 
101
102
103
104
105
106
run:
        with open(input[0], 'r') as f:
                x= f.readlines()
        x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
        with open(output[0], 'w') as f:
                f.write(''.join(x))
SnakeMake From line 101 of LDscore_reprpheno/Snakefile 
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
run:
	allfiles= [infile for infile in input if wildcards.repr_pheno not in infile]
	allfiles= ','.join(allfiles)
	outfile= params[0] + wildcards.repr_pheno + '_rg'
	infile= input[0]
	shell("""
	set +eu
	source /home/pol/miniconda3/etc/profile.d/conda.sh
	conda activate ldsc
	python2 /home/pol/software/ldsc/ldsc.py \
	--rg {infile},{allfiles} \
	--ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
	--w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
	--out {outfile}
	conda deactivate
	set -eu
	""")
SnakeMake From line 117 of LDscore_reprpheno/Snakefile 
141
142
143
144
145
146
run:
        with open(input[0], 'r') as f:
                x= f.readlines()
        x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
        with open(output[0], 'w') as f:
                f.write(''.join(x))
SnakeMake From line 141 of LDscore_reprpheno/Snakefile 
154
155
156
157
158
159
160
run:
	df_list= list()
	for i in input:
		d= pd.read_csv(i, delim_whitespace= True, header= 0)
		df_list.append(d)
	d= pd.concat(df_list)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 154 of LDscore_reprpheno/Snakefile 
172
173
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0)
SnakeMake From line 172 of LDscore_reprpheno/Snakefile 
211
212
213
214
215
216
217
218
219
220
221
222
223
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/munge_sumstats.py \
        --out {params[0]} \
        --merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
        --sumstats {input[0]} \
        --chunksize 500000
        conda deactivate
        set -eu
        """
SnakeMake From line 211 of LDscore_reprpheno/Snakefile 
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
run:
        allfiles= [infile for infile in input if 'BW_maternal_effect' not in infile]
        allfiles= ','.join(allfiles)
        outfile= params[0] + 'BW_maternal_effect_rg'
        infile= input[0]
        shell("""
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/ldsc.py \
        --rg {infile},{allfiles} \
        --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --out {outfile}
        conda deactivate
        set -eu
        """)
SnakeMake From line 234 of LDscore_reprpheno/Snakefile 
261
262
263
264
265
266
267
268
269
270
271
272
273
	shell:
		"""
		set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2 {input[0]} \
                --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --out {params[0]}
                conda deactivate
                set -eu
		"""
SnakeMake From line 261 of LDscore_reprpheno/Snakefile 
281
282
283
284
285
286
run:
        with open(input[0], 'r') as f:
                x= f.readlines()
        x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
        with open(output[0], 'w') as f:
                f.write(''.join(x))
SnakeMake From line 281 of LDscore_reprpheno/Snakefile 
297
298
299
300
301
302
303
304
305
306
307
308
309
	shell:
		"""
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2 {input[0]} \
                --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --out {params[0]}
                conda deactivate
                set -eu
                """
SnakeMake From line 297 of LDscore_reprpheno/Snakefile 
317
318
319
320
321
322
323
324
325
326
327
328
	run:
                df_list= list()
                for infile in input:
                        with open(infile, 'r') as f:
                                lines= [line.strip() for line in f if line.startswith('Total Observed')]
                                h2= float(lines[0].split(' ')[4])
                                se= float(lines[0].split('(')[1].replace(')', ''))
                                cohort= infile.split('/')[10].replace('_h2.log', '')
                                d= pd.DataFrame({'cohort': cohort, 'h2': h2, 'se': se}, index= [0])
                                df_list.append(d)
                d= pd.concat(df_list)
                d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 317 of LDscore_reprpheno/Snakefile 
17
18
19
20
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0, compression= 'gzip', usecols= ['RSID', 'CHR', 'POS', 'TOTALSAMPLESIZE', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])
        d.columns= ['CHR', 'POS', 'A1', 'A2', 'N', 'BETA', 'SE', 'pvalue', 'SNP']
        d.dropna(axis= 0, inplace= True)
SnakeMake From line 17 of LDscore/Snakefile 
40
41
42
43
44
45
46
47
48
49
50
51
52
        shell:
                """
		set +eu
		source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/munge_sumstats.py \
		--merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
                --out {params[0]} \
                --sumstats {input[0]} \
		--chunksize 500000
                conda deactivate
		set -eu
                """
SnakeMake From line 40 of LDscore/Snakefile 
63
64
65
run:
	allfiles= [infile for infile in input if wildcards.pheno not in infile]
	allfiles= ','.join(allfiles)
SnakeMake From line 63 of LDscore/Snakefile 
88
89
90
91
92
93
run:
	with open(input[0], 'r') as f:
		x= f.readlines()
	x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
	with open(output[0], 'w') as f:
		f.write(''.join(x))
SnakeMake From line 88 of LDscore/Snakefile 
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
	shell:
		'''
		set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
		python2 /home/pol/software/ldsc/ldsc.py \
		--h2 {input[0]}\
		--ref-ld-chr /home/pol/software/ldsc/baseline/baseline/baselineLD. \
		--w-ld-chr /home/pol/software/ldsc/baseline/weights_hm3_no_hla/weights.\
		--overlap-annot\
		--frqfile-chr /home/pol/software/ldsc/baseline/1000G_Phase3_frq/1000G.EUR.QC.\
		--out {params[0]}
		conda deactivate
                set -eu
		'''
SnakeMake From line 104 of LDscore/Snakefile 
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
        shell:
                '''
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
		cd /home/pol/software/ldsc/cts/
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2-cts {input[0]}\
                --ref-ld-chr-cts {params[1]} \
                --w-ld-chr /home/pol/software/ldsc/baseline/weights_hm3_no_hla/weights.\
		--ref-ld-chr /home/pol/software/ldsc/baseline/baseline/baselineLD.\
                --out {params[0]}
                conda deactivate
                set -eu
                '''
SnakeMake From line 130 of LDscore/Snakefile 
156
157
	run:
                d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['SNP', 'CHR', 'POS', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])[['SNP', 'CHR', 'POS', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']]
SnakeMake From line 156 of LDscore/Snakefile 
185
186
187
188
189
190
191
192
193
194
195
196
197
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/munge_sumstats.py \
        --merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
        --out {params[0]} \
        --sumstats {input[0]} \
        --chunksize 500000
        conda deactivate
        set -eu
        """
SnakeMake From line 185 of LDscore/Snakefile 
209
210
211
212
213
214
215
216
217
218
219
220
221
	shell:
                """
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2 {input[0]} \
                --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --out {params[0]}
                conda deactivate
                set -eu
                """
SnakeMake From line 209 of LDscore/Snakefile 
229
230
231
232
233
234
235
236
237
238
239
240
run:
        df_list= list()
        for infile in input:
                with open(infile, 'r') as f:
                        lines= [line.strip() for line in f if line.startswith('Total Observed')]
                        h2= float(lines[0].split(' ')[4])
                        se= float(lines[0].split('(')[1].replace(')', ''))
                        cohort= infile.split('/')[9].replace('_h2.log', '')
                        d= pd.DataFrame({'cohort': cohort, 'h2': h2, 'se': se}, index= [0])
                        df_list.append(d)
        d= pd.concat(df_list)
        d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 229 of LDscore/Snakefile 
252
253
254
255
256
257
258
259
	run:
                d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['SNP', 'CHR', 'POS', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue'])[['SNP', 'CHR', 'POS', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue']]
                d['SNP']= d.SNP.str.replace(':SNP', '')
                d['SNP']= d.SNP.str.replace(':INDEL', '')
                d['CHR']= d.CHR.apply(str)
                d.columns= ['ID', 'CHR', 'POS', 'N', 'A2', 'A1', 'BETA', 'SE', 'pvalue']
                d.dropna(axis= 0, inplace= True)
                d['CHR']= d.CHR.apply(str)
SnakeMake From line 252 of LDscore/Snakefile 
281
282
283
284
285
286
287
288
289
290
291
292
293
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/munge_sumstats.py \
        --merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
        --out {params[0]} \
        --sumstats {input[0]} \
        --chunksize 500000
        conda deactivate
        set -eu
        """
SnakeMake From line 281 of LDscore/Snakefile 
304
305
306
307
308
309
310
311
312
313
314
315
316
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/ldsc.py \
        --h2 {input[0]} \
        --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --out {params[0]}
        conda deactivate
        set -eu
        """
SnakeMake From line 304 of LDscore/Snakefile 
324
325
326
327
328
329
330
331
332
333
334
335
run:
	df_list= list()
	for infile in input:
		with open(infile, 'r') as f:
			lines= [line.strip() for line in f if line.startswith('Total Observed')]
			h2= float(lines[0].split(' ')[4])
			se= float(lines[0].split('(')[1].replace(')', ''))
			cohort= infile.split('/')[10].replace('_allPTD.log', '')
			d= pd.DataFrame({'cohort': cohort, 'h2': h2, 'se': se}, index= [0])
			df_list.append(d)
	d= pd.concat(df_list)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 324 of LDscore/Snakefile 
345
346
347
348
349
350
351
352
353
354
355
356
357
	shell:
                """
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2 {input[0]} \
                --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
                --out {params[0]}
                conda deactivate
                set -eu
                """
SnakeMake From line 345 of LDscore/Snakefile 
365
366
run:
	df_list= list()
SnakeMake From line 365 of LDscore/Snakefile 
387
388
389
390
391
run:
	x= pd.read_csv(input[0], sep= '\t', header= 0)
	d= pd.read_csv(input[1], sep= '\t', header= 0)
	d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
	d[['CHR', 'POS', 'REF', 'EFF', 'INDELS']]= d['MarkerName'].str.split(':', expand= True)
SnakeMake From line 387 of LDscore/Snakefile 
416
417
418
419
420
421
422
423
424
425
426
427
428
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/munge_sumstats.py \
        --merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
        --out {params[0]} \
        --sumstats {input[0]} \
        --chunksize 500000
        conda deactivate
        set -eu
        """
SnakeMake From line 416 of LDscore/Snakefile 
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
run:
        allfiles= [infile for infile in input if wildcards.PTD_metas not in infile]
        allfiles= ','.join(allfiles)
        outfile= params[0] + wildcards.PTD_metas + '_rg'
        infile= input[0]
        shell("""
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/ldsc.py \
        --rg {infile},{allfiles} \
        --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --out {outfile}
        conda deactivate
        set -eu
        """)
SnakeMake From line 440 of LDscore/Snakefile 
464
465
466
467
468
469
470
471
472
473
474
run:
	with open(input[0], 'r') as f:
		x= f.readlines()
	x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
	with open(output[0], 'w') as f:
		f.write(''.join(x))
	with open(input[1], 'r') as f:
		x= f.readlines()
	x= x[x.index('Summary of Genetic Correlation Results\n')+2:-3]
	with open(output[0], 'a') as f:
		f.write(''.join(x))
SnakeMake From line 464 of LDscore/Snakefile 
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
run:
	allfiles= [infile for infile in input if 'individual_cohorts/' + wildcards.big5 not in infile]
	allfiles= ','.join(allfiles)
	print(allfiles)
	outfile= input[0].replace('.txt.sumstats.gz', '_rg')
	infile= input[0]
	shell("""
	set +eu
	source /home/pol/miniconda3/etc/profile.d/conda.sh
	conda activate ldsc
	python2 /home/pol/software/ldsc/ldsc.py \
	--rg {infile},{allfiles} \
	--ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
	--w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
	--out {params[0]}
	conda deactivate
	set -eu
	""")
SnakeMake From line 488 of LDscore/Snakefile 
513
514
515
516
517
518
519
520
521
522
523
524
run:
	for i in range(len(input)):
		with open(input[i], 'r') as f:
			x= f.readlines()
		if i== 0: 
			x= x[x.index('Summary of Genetic Correlation Results\n')+1:-3]
			with open(output[0], 'w') as f:
				f.write(''.join(x))
		else:
			x= x[x.index('Summary of Genetic Correlation Results\n')+2:-3]
			with open(output[0], 'a') as f:
				f.write(''.join(x))
SnakeMake From line 513 of LDscore/Snakefile 
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d['Allele1']= d['Allele1'].str.upper()
	d['Allele2']= d['Allele2'].str.upper()
	d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
	d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
	d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
	d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
	d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d['pvalue']= d['pvalue'].astype(str).astype(float)
	d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
	d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
	d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
	rs= pd.read_csv(input[1], sep= '\t', header=0)
	rs.columns= ['ID', 'RSID']
SnakeMake From line 533 of LDscore/Snakefile 
564
565
566
567
568
569
570
571
572
573
574
575
576
run:
        x= pd.read_csv(input[0], sep= '\t', header= 0)
        d= pd.read_csv(input[1], sep= '\t', header= 0)
        d['CHR']= np.where(d['CHR']== 'X', '23', d['CHR'])
        d['POS']= d['POS'].astype(str).astype(int)
        d['CHR']= d['CHR'].astype(str).astype(int)
        d.dropna(axis= 0, inplace= True)
        d= pd.merge(d, x[['CHR', 'SNP', 'BP']], left_on= ['CHR', 'POS'], right_on= ['CHR', 'BP'])
        d= d.loc[~((d.CHR==6) & (d.POS >28477797) & (d.POS< 33448354)), :]
        d= d[['CHR', 'POS', 'RSID', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']]
        d.columns= ['CHR', 'POS', 'SNP', 'A1', 'A2', 'N', 'EAF', 'BETA', 'SE', 'pvalue']
        d.drop_duplicates(['CHR', 'POS', 'A1', 'A2'], keep= 'first', inplace= True)
        d.to_csv(output[0], sep= '\t', header= True, index= False, columns= ['SNP', 'CHR', 'POS', 'N', 'A2', 'A1', 'BETA', 'SE', 'pvalue'])
SnakeMake From line 564 of LDscore/Snakefile 
588
589
590
591
592
593
594
595
596
597
598
599
600
shell:
        """
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/munge_sumstats.py \
        --merge-alleles /home/pol/software/ldsc/w_hm3.snplist \
        --out {params[0]} \
        --sumstats {input[0]} \
        --chunksize 500000
        conda deactivate
        set -eu
        """
SnakeMake From line 588 of LDscore/Snakefile 
611
612
613
614
615
616
617
618
619
620
621
622
623
run:
        shell("""
        set +eu
        source /home/pol/miniconda3/etc/profile.d/conda.sh
        conda activate ldsc
        python2 /home/pol/software/ldsc/ldsc.py \
        --rg {input[0]},{input[1]} \
        --ref-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --w-ld-chr /home/pol/software/ldsc/eur_w_ld_chr/ \
        --out {params[0]}
        conda deactivate
        set -eu
        """)
SnakeMake From line 611 of LDscore/Snakefile 
637
638
639
640
641
642
643
644
run:
	d= pd.read_table(input[0], sep= '\t', header= 0)
	d['kbid']= d.kbid.str.split('.', expand= True)[0]
	d['Cell_type']= d.Cell_type.str.replace(' ', '-')
	for k, g in d[d['Cell_type'].isin(set(d.Cell_type))].groupby('Cell_type'):
		g.to_csv(params[0] + k + '.txt', header= False, sep= '\t', columns= ['kbid'], index= False)
	d.drop_duplicates('kbid', inplace= True, keep= 'first')
	d.to_csv(output[-1], sep= '\t', header= False, index= False, columns= ['kbid'])
SnakeMake From line 637 of LDscore/Snakefile 
655
656
657
run:
	shell("""
	set +eu
SnakeMake From line 655 of LDscore/Snakefile 
681
682
683
run:
	shell("""
	set +eu
SnakeMake From line 681 of LDscore/Snakefile 
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
	shell:
                '''
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2 {input[0]}\
                --ref-ld-chr /home/pol/software/ldsc/baseline/baseline/baselineLD.,{params[1]}. \
                --w-ld-chr /home/pol/software/ldsc/baseline/weights_hm3_no_hla/weights.\
                --overlap-annot \
                --frqfile-chr /home/pol/software/ldsc/baseline/1000G_Phase3_frq/1000G.EUR.QC. \
                --out {params[0]} \
		--thin-annot
                conda deactivate
                set -eu
                '''
SnakeMake From line 712 of LDscore/Snakefile 
736
737
738
739
740
741
742
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d= d.loc[d.Category== 'L2_1', :]
	x= pd.read_csv(input[1], sep= '\t', header= None, names= ['Gene'])
	d['n_genes']= x.shape[0]
	d['Category']= wildcards.cell_types
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 736 of LDscore/Snakefile 
750
751
752
753
754
shell:
	'''
	head -1 {input[0]} > {output[0]}
	tail -n +2 -q {input} >> {output[0]}
	'''
SnakeMake From line 750 of LDscore/Snakefile 
766
767
run:
	shell("""
SnakeMake From line 766 of LDscore/Snakefile 
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
	run:

                shell("""
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --l2 \
                --bfile {params[0]} \
                --ld-wind-cm 1 \
                --annot {input[0]} \
                --out {params[1]} \
                --print-snps {input[1]} \
                --thin-annot
                """)
SnakeMake From line 794 of LDscore/Snakefile 
823
824
825
826
827
828
run:
        if wildcards.cell_types!= 'overall':
                d= pd.DataFrame({'V1': [wildcards.cell_types], 'V2': ','.join(params)})
                d.to_csv(output[0], sep= '\t', header= False, index= False)
        else:
                open(output[0], 'a').close()
SnakeMake From line 823 of LDscore/Snakefile 
836
837
shell:
        'cat {input} > {output[0]}'
SnakeMake From line 836 of LDscore/Snakefile 
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
        shell:
                '''
                set +eu
                source /home/pol/miniconda3/etc/profile.d/conda.sh
                conda activate ldsc
                python2 /home/pol/software/ldsc/ldsc.py \
                --h2-cts {input[0]}\
                --ref-ld-chr /home/pol/software/ldsc/baseline/baseline/baselineLD. \
		--ref-ld-chr-cts {input[1]} \
                --w-ld-chr /home/pol/software/ldsc/baseline/weights_hm3_no_hla/weights.\
                --overlap-annot \
                --frqfile-chr /home/pol/software/ldsc/baseline/1000G_Phase3_frq/1000G.EUR.QC. \
                --out {params[0]} \
                --thin-annot
                conda deactivate
                set -eu
                '''
SnakeMake From line 855 of LDscore/Snakefile 
10
11
shell:
	'grep -v {wildcards.allPTD_coh} {input[0]} | sed -e "s/to_replace/{params[0]}/g" > {output[0]}'
SnakeMake From line 10 of LOCO_meta_allPTD/Snakefile 
38
39
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['MarkerName', 'Allele1', 'P-value'])
SnakeMake From line 38 of LOCO_meta_allPTD/Snakefile 
59
60
shell:
        'bedtools closest -t all -a {input[0]} -b {input[1]} > {output[0]}'
SnakeMake BEDTools From line 59 of LOCO_meta_allPTD/Snakefile 
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
run:
	d= pd.read_csv(input[0], sep= '\t', header=0)
	d['Allele1']= d['Allele1'].str.upper()
	d['Allele2']= d['Allele2'].str.upper()
	d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
	d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value', 'HetISq', 'HetPVal']]
	d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue', 'HetISq', 'HetPVal']
	d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
	d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d['pvalue']= d['pvalue'].astype(str).astype(float)
	d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
	d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
	d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
SnakeMake From line 69 of LOCO_meta_allPTD/Snakefile 
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
run:
        d= pd.read_csv(input[0], sep= '\t', compression= 'gzip', usecols= ['CHR', 'POS', 'pvalue', 'nearestGene', 'ID'])
        df= d.loc[d.pvalue< 5*10**-8, :]
        df.sort_values(by= 'pvalue', ascending= True, inplace= True)
        df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
        df_list= list()
        for chrom in set(df.CHR):
                d_temp= df.loc[df.CHR== chrom, :]
                positions= d_temp.POS.values
                for pos in positions:
                        if pos in d_temp.POS.values:
                                df_list.append(d_temp.loc[d_temp.POS== pos, :])
                                d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
                        else:
                                continue
        x= pd.concat(df_list)
        x['pos1']= x.POS - 1.5*10**6
        x['pos2']= x.POS + 1.5*10**6
        x['CHR']= x.CHR.astype(str)
        x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
        x.to_csv(output[0], sep='\t', header= True, index= False, columns= ['CHR', 'pos1', 'pos2', 'nearestGene', 'ID', 'pvalue'])
SnakeMake From line 98 of LOCO_meta_allPTD/Snakefile 
10
11
shell:
	'grep -v {wildcards.GAraw_coh} {input[0]} | sed -e "s/to_replace/{params[0]}/g" > {output[0]}'
SnakeMake From line 10 of LOCO_meta/Snakefile 
38
39
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['MarkerName', 'Allele1', 'P-value'])
SnakeMake From line 38 of LOCO_meta/Snakefile 
59
60
shell:
        'bedtools closest -t all -a {input[0]} -b {input[1]} > {output[0]}'
SnakeMake BEDTools From line 59 of LOCO_meta/Snakefile 
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
run:
	d= pd.read_csv(input[0], sep= '\t', header=0)
	d['Allele1']= d['Allele1'].str.upper()
	d['Allele2']= d['Allele2'].str.upper()
	d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
	d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value', 'HetISq', 'HetPVal']]
	d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue', 'HetISq', 'HetPVal']
	d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
	d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d['pvalue']= d['pvalue'].astype(str).astype(float)
	d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
	d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
	d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
SnakeMake From line 69 of LOCO_meta/Snakefile 
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
run:
        d= pd.read_csv(input[0], sep= '\t', compression= 'gzip', usecols= ['CHR', 'POS', 'pvalue', 'nearestGene', 'ID'])
        df= d.loc[d.pvalue< 5*10**-8, :]
        df.sort_values(by= 'pvalue', ascending= True, inplace= True)
        df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
        df_list= list()
        for chrom in set(df.CHR):
                d_temp= df.loc[df.CHR== chrom, :]
                positions= d_temp.POS.values
                for pos in positions:
                        if pos in d_temp.POS.values:
                                df_list.append(d_temp.loc[d_temp.POS== pos, :])
                                d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
                        else:
                                continue
        x= pd.concat(df_list)
        x['pos1']= x.POS - 1.5*10**6
        x['pos2']= x.POS + 1.5*10**6
        x['CHR']= x.CHR.astype(str)
        x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
        x.to_csv(output[0], sep='\t', header= True, index= False, columns= ['CHR', 'pos1', 'pos2', 'nearestGene', 'ID', 'pvalue'])
SnakeMake From line 98 of LOCO_meta/Snakefile 
11
12
13
14
15
16
17
run:
	d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['CHR', 'POS', 'pvalue', 'nearestGene'], compression= 'gzip')
	d.sort_values(['pvalue'], ascending=True, inplace= True)
	d.drop_duplicates(['CHR', 'POS'], inplace= True, keep= 'first')
	d['ID']= 'chr' + d.CHR.apply(str) + ':' + d.POS.apply(str)
	d.columns= ['CHR', 'POS', 'P-value', 'nearestGene', 'MarkerName']
	d.to_csv(output[0], sep= '\t', header= True, index= False, columns= ['MarkerName', 'P-value'])
SnakeMake From line 11 of LocusZoom/Snakefile 
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
run:
	df= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['CHR', 'POS', 'pvalue', 'nearestGene'], compression= 'gzip')
	df= df.loc[df.pvalue< 5*10**-8, :]
	df.sort_values(by= 'pvalue', ascending= True, inplace= True)
	df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
	df_list= list()
	for chrom in set(df.CHR):
		d_temp= df.loc[df.CHR== chrom, :]
		positions= d_temp.POS.values
		for pos in positions:
			if pos in d_temp.POS.values:
				df_list.append(d_temp.loc[d_temp.POS== pos, :])
				d_temp= d_temp.loc[(d_temp.POS < pos - (1.5 * 10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
			else:
				continue
	df= pd.concat(df_list)
	df['CHR']= df.CHR.astype(str)
SnakeMake From line 25 of LocusZoom/Snakefile 
57
58
59
60
61
62
63
64
65
66
67
run:
	if not os.path.exists(params[1]):
		os.makedirs(params[1])
	df= pd.read_csv(input[0], sep= '\t', header= 0)
	for index, row in df.iterrows():
		snp= row['snp']
		title= '"' + row['nearestGene'] + '"'
		shell('python2 /home/pol/software/locuszoom/bin/locuszoom --metal {input[1]} --refsnp {snp} --flank 250kb --plotonly --no-date --build hg19 --pop EUR --source 1000G_March2012 --prefix {params[0]} title={title} theme=publication')
		outfile= params[1] + 'chr' + str(row['chr']) + '_' + row['nearestGene'] + '.pdf'
		infile= params[0] + '_' + snp.replace(':', '_') + '.pdf'
		shell('qpdf --empty --pages {infile} 1 -- {outfile}; rm {infile}')
SnakeMake From line 57 of LocusZoom/Snakefile 
80
81
82
run:
	if len(input)== 1:
		shell('cp {input[0]} {output[0]}')
SnakeMake From line 80 of LocusZoom/Snakefile 
96
97
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['CHR', 'POS', 'pvalue'], compression= 'gzip')
SnakeMake From line 96 of LocusZoom/Snakefile 
126
127
128
129
run:
	if not os.path.exists(params[1]):
		os.makedirs(params[1])
	shell('python2 /home/pol/software/locuszoom/bin/locuszoom --metal {input[0]} --refsnp rs9823520 --flank 250kb --plotonly --no-date --build hg19 --pop EUR --source 1000G_March2012 --prefix {params[0]} title="WNT4-GA" theme=publication')
SnakeMake From line 126 of LocusZoom/Snakefile 
20
21
shell:
	'/home/pol/software/generic-metal/metal {params[0]} >> {output[1]}'
SnakeMake From line 20 of meta/Snakefile 
37
38
39
40
41
42
    shell:
        '''
        /home/pol/software/generic-metal/metal {input[0]} >> {output[3]}
        /home/pol/software/generic-metal/metal {input[1]} >> {output[3]}
	/home/pol/software/generic-metal/metal {input[2]} >> {output[3]}
        '''
SnakeMake From line 37 of meta/Snakefile 
55
56
57
58
59
shell:
    '''
    /home/pol/software/generic-metal/metal {input[0]} >> {output[2]}
    /home/pol/software/generic-metal/metal {input[1]} >> {output[2]}
    '''
SnakeMake From line 55 of meta/Snakefile 
70
71
shell:
        '/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 70 of meta/Snakefile 
81
82
shell:
        '/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 81 of meta/Snakefile 
92
93
shell:
        '/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 92 of meta/Snakefile 
104
105
shell:
        '/home/pol/software/generic-metal/metal {input[0]} >> {output[1]}'
SnakeMake From line 104 of meta/Snakefile 
115
116
117
118
shell:
    '''
    /home/pol/software/generic-metal/metal {input[0]} >> {output[1]}
    '''
SnakeMake From line 115 of meta/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
library(MendelianRandomization)
library(data.table)
library(dplyr)

if (!grepl('cluster', snakemake@output[[1]])){
d= fread(snakemake@input[[1]])
names(d)= c('ID', 'beta', 'se', 'pvalue', 'trait')
} else {
d= fread(snakemake@input[[1]])

}
x=fread(snakemake@input[[2]])
x= filter(x, !duplicated(ID))
d= inner_join(d, x, by= 'ID')



funk= function(temp_df){

inputMR= mr_input(bx = temp_df$beta,   bxse= temp_df$se,by = temp_df$BETA, byse = temp_df$SE)

if (nrow(temp_df)>3) {

z= mr_allmethods(inputMR)$Values
names(z)= c('method', 'estimate', 'se', 'lo95', 'up95', 'pvalue')
z$trait= unique(temp_df$trait)

} else {
z= mr_ivw(inputMR)

z= data.frame(method= 'IVW', estimate= z$Estimate, se= z$StdError, lo95= z$CILower, up95= z$CIUpper, pvalue= z$Pvalue, trait= unique(temp_df$trait))

}
return(z)
}


mr= lapply(split(d, d$trait), funk)

mr= do.call('rbind', mr)

fwrite(mr, snakemake@output[[1]], sep= '\t')
R dplyr data.table From line 1 of MR/MR_reproductive_traits.R 
16
17
18
19
20
21
22
23
24
25
run:
	d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['CHR', 'POS', 'pvalue', 'ID'])
	x= pd.read_csv(input[1], sep= '\t', header= 0, usecols= ['ID', 'EAF'])
	x= x.loc[((x.EAF>=0.01) & (x.EAF<= 0.99)), :]
	d= d.loc[d.pvalue< 5e-8, :]
	d= d.loc[d.ID.isin(x.ID.values), :]
	d.drop_duplicates('ID', inplace= True)
	if d.shape[0] == 0: 
		open(output[0], 'a').close()
	else:
SnakeMake From line 16 of MR/Snakefile 
40
41
42
43
44
45
46
47
48
49
50
run:
	d= pd.read_csv(input[0], sep= '\t', header= None, names= ['CHR', 'SNP', 'x1', 'POS', 'A1', 'A2'])
	d['A1']= np.where(d.A1.str.len() > d.A2.str.len(), 'I', d.A1)
	d['A2']= np.where(d.A1.str.len() < d.A2.str.len(), 'I', d.A2)
	d['A1']= np.where(d.A2== 'I', 'D', d.A1)
	d['A2']= np.where(d.A1== 'I', 'D', d.A2)
	d['SNP']= np.where(d.A1>d.A2, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.A2 + ':' + d.A1, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.A1 + ':' + d.A2)
	d.to_csv(output[0], sep= '\t', header= False, index= False)
	d= d[d.duplicated(['SNP'], keep= False)]
	d.drop_duplicates('SNP', inplace= True, keep= 'first')
	d.to_csv(output[1], sep='\t', columns= ['SNP'], index= False, header= False)
SnakeMake From line 40 of MR/Snakefile 
64
65
run:
        shell('~/software/plink --bim {input[2]} --bed {input[3]} --fam {input[4]} --clump {input[0]} --exclude {input[1]} --clump-r2 0.001 --clump-kb 1000 --clump-p1 5e-8 --clump-p2 1e-5 --out {params[1]} || true')
SnakeMake pLink From line 64 of MR/Snakefile 
75
76
77
78
79
80
81
82
83
84
85
run:
	if os.stat(input[1]).st_size == 0:
		open(output[0], "w").close
	else:
		d= pd.read_csv(input[0], sep='\t', header= 0, usecols= ['ID', 'BETA', 'SE', 'pvalue'])
		x= pd.read_csv(input[1], delim_whitespace= True, header= 0)
		d= d.loc[d.ID.isin(x.SNP.values), :]
		d= d.groupby('ID').head(1)
		d= d[['ID', 'BETA', 'SE', 'pvalue']]
		d['trait']= wildcards.repr_pheno
		d.to_csv(output[0], sep= '\t', header= False, index= False)
SnakeMake From line 75 of MR/Snakefile 
93
94
shell:
	'echo -e "ID\tbeta\tse\tpvalue_exp\ttrait" | cat {input} > {output[0]}'
SnakeMake From line 93 of MR/Snakefile 
104
105
script:
        'MR_reproductive_traits.R'
SnakeMake From line 104 of MR/Snakefile 
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0)
        x= pd.read_csv(input[1], sep= '\t', header= 0, usecols= ['#CHROM', 'POS', 'ID', 'REF', 'ALT'])
        x.columns= ['CHR', 'POS', 'ID', 'REF', 'ALT']
        df= d.loc[(d.OKG_proxy != 'Signal in 1KG') & (d.OKG_proxy != 'No 1KG proxy'), :]
        okg= df.loc[(df.OKG_Other_allele.str.len() == 1) & (d.OKG_Trait_raising.str.len() == 1), :]
        hm= df.loc[(df.OKG_Other_allele.str.len() != 1) | (d.OKG_Trait_raising.str.len() != 1), :]
        hm['beta']= hm.HM_Weight
        hm['ref']= hm.HM_Other_allele
        hm['eff']= hm.HM_Trait_raising
        hm['RSID']= hm.HM_proxy
        hm['se']= hm.HM_SE_weight
        hm= hm[['RSID', 'beta', 'se', 'ref', 'eff', 'Cluster']]
        okg['beta']= okg.OKG_Weight
        okg['ref']= okg.OKG_Other_allele
        okg['eff']= okg.OKG_Trait_raising
        okg['RSID']= okg.OKG_proxy
        okg['se']= okg.OKG_SE_weight
        okg= okg[['RSID', 'beta', 'se', 'ref', 'eff', 'Cluster']]
        d= d.loc[(d.OKG_proxy == 'Signal in 1KG') | (d.OKG_proxy == 'No 1KG proxy'), :]
        d['beta']= d.Weight
        d['ref']= d.Other_allele
        d['eff']= d.Trait_raising
        d['RSID']= d.Signal
        d['se']= d.SE_weight
        d= d[['RSID', 'beta', 'se', 'ref', 'eff', 'Cluster']]
        d= pd.concat([d, hm, okg])
        d= pd.merge(d, x, left_on= ['RSID'], right_on= 'ID')
        d= d.loc[(d.ALT== d.ref) | (d.REF== d.ref), :]
        d= d.loc[(d.ALT== d.eff) | (d.REF== d.eff), :]
        d['beta']= np.where(d.ref > d.eff, -1 * d.beta, d.beta)
        d['ID']= np.where(d.ref > d.eff, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.eff + ':' + d.ref, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.ref + ':' + d.eff)
        d['trait']= np.where(d.Cluster== 'Female SHBG cluster', 'SHBG_fem_cluster', 'Testosterone_fem_cluster')
        d.to_csv(output[0], sep= '\t', header= True, index= False, columns= ['ID', 'beta', 'se', 'trait'])
SnakeMake From line 115 of MR/Snakefile 
158
159
script:
        'MR_reproductive_traits.R'
SnakeMake From line 158 of MR/Snakefile 
168
169
170
171
172
173
174
run:
	x= pd.read_csv(input[0], sep= '\t', header= None, names= ['ID', 'beta', 'se', 'pvalue', 'trait'])
	x= x.loc[((x.trait== 'SHBG_fem') | (x.trait== 'Testosterone_fem') | (x.trait== 'CBAT_fem')), :]
	x.drop_duplicates(subset= 'ID', inplace= True)
	x[['CHR', 'POS', 'REF', 'EFF']]= x.ID.str.split(':', expand= True)
	x['CHR']= np.where(x.CHR== 'X', '23', x.CHR)
	x.to_csv(output[0], sep= '\t', header= False, index= False, columns= ['CHR', 'POS', 'POS', 'ID'])
SnakeMake From line 168 of MR/Snakefile 
186
187
shell:
	'~/software/plink2 --bfile {params[0]} --extract bed1 {input[0]} --memory 5000 --threads {threads} --make-bed --out {params[1]}'
SnakeMake plink2 From line 186 of MR/Snakefile 
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
run:
        d= pd.read_csv(input[0], sep= '\t', header= None, names= ['CHR', 'SNP', 'x1', 'POS', 'A1', 'A2'])
        d['A1']= np.where(d.A1.str.len() > d.A2.str.len(), 'I', d.A1)
        d['A2']= np.where(d.A1.str.len() < d.A2.str.len(), 'I', d.A2)
        d['A1']= np.where(d.A2== 'I', 'D', d.A1)
        d['A2']= np.where(d.A1== 'I', 'D', d.A2)
        d['CHR']= d.CHR.apply(str)
        d['CHR']= np.where(d.CHR== 'X', '23', d.CHR)
        d['SNP']= np.where(d.A1>d.A2, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.A2 + ':' + d.A1, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.A1 + ':' + d.A2)
        d.to_csv(output[0], sep= '\t', header= False, index= False)
        d= d[d.duplicated(['SNP'], keep= False)]
        d.drop_duplicates('SNP', inplace= True, keep= 'first')
        d.to_csv(output[3], sep='\t', columns= ['SNP'])
        shell('mv {input[1]} {output[1]}')
        shell('mv {input[2]} {output[2]}')
SnakeMake From line 200 of MR/Snakefile 
225
226
shell:
        '~/software/plink --bfile {params[0]} --r square --out {params[1]}'
SnakeMake pLink From line 225 of MR/Snakefile 
240
241
script:
	'MVMR.R'
SnakeMake From line 240 of MR/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
library(data.table)
library(dplyr)


hrc= fread(snakemake@input[[1]])


funk= function(infile){
d= fread(infile)

print(paste('Filtering file: ', infile))

d= arrange(d, CHR, POS, EFF, REF)

d= filter(d, pvalue< 1, pvalue>0)

d$pval= pnorm(-abs(d$BETA / d$SE)) * 2

d= filter(d, (abs(-log10(pvalue) - -log10(pval)) / -log10(pval)) * 100 <= 10)

d$ID= with(d, ifelse(REF> EFF, paste(CHR, POS, EFF, REF, sep= ':'), paste(CHR, POS, REF, EFF, sep= ':')))

d$SNP= with(d, ifelse(grepl('I', ID), paste(ID, 'INDEL', sep= ':'), paste(ID, 'SNP', sep= ':')))


print(str(d))

d= inner_join(d, hrc, by= 'ID')
d$EAF= ifelse(is.na(d$EAF), d$eaf, d$EAF)

d$BETA= ifelse(d$REF> d$EFF, -1 * d$BETA, d$BETA)
d$EAF= ifelse(d$REF> d$EFF, 1 - d$EAF, d$EAF)

d[d$REF>d$EFF, c("REF", "EFF")]= d[d$REF > d$EFF, c("EFF", "REF")]

d$MAF= ifelse(d$EAF>0.5, 1- d$EAF, d$EAF)
d= filter(d, MAF>= 0.005)

d= filter(d, pvalue>0, pvalue<1, MAF>=0.005, SE>0)

d= filter(d, (MAF * 2 * N) > 6)

d$maf= ifelse(d$eaf> 0.5, 1 - d$eaf, d$eaf)

d= filter(d, abs(maf - MAF) < 0.2)

if (grepl('GAraw/Viva', infile)){



d$EAF= with(d, ifelse(abs(eaf - EAF)> 0.2, 1 - EAF, EAF))
d$BETA= with(d, ifelse(abs(eaf - EAF)> 0.2, -1 * BETA, BETA))
}


if (grepl('GAnrm/Viva', infile)){



d$EAF= with(d, ifelse(abs(eaf - EAF)> 0.2, 1 - EAF, EAF))
d$BETA= with(d, ifelse(abs(eaf - EAF)> 0.2, -1 * BETA, BETA))
}

if (grepl('postTerm/HUNT', infile)){



d$EAF= with(d, ifelse(abs(eaf - EAF)> 0.2, 1 - EAF, EAF))
d$BETA= with(d, ifelse(abs(eaf - EAF)> 0.2, -1 * BETA, BETA))
}


d= arrange(d, pvalue)
d= filter(d, !duplicated(ID))


d= select(d, -c(MAF, ID, eaf, pval))

x2= nrow(d)

d$STRAND= '+'

#outfile= paste0(snakemake@params[[1]], gsub('_temp.txt', '', unlist(strsplit(infile, '/'))[9]), '.txt')


fwrite(d, snakemake@output[[1]], sep= '\t')
}


#input_files= snakemake@input[grepl('sumstats', snakemake@input)]

lapply(snakemake@input[[2]], funk)
R dplyr data.table From line 1 of munge_stats/filter_SNPs.R 
96
97
run:
    format_list(input[0], output[0])
SnakeMake From line 96 of munge_stats/Snakefile 
107
108
109
110
run:
	for infile in input:
		outfile= params[0] + infile.split('-')[0].split('/')[-1] + '_temp.txt'
		shell('mv {infile} {outfile}')
SnakeMake From line 107 of munge_stats/Snakefile 
118
119
run:
    format_list(input[0], output[0])
SnakeMake From line 118 of munge_stats/Snakefile 
129
130
131
132
run:
        for infile in input:
                outfile= params[0] + infile.split('-')[0].split('/')[-1] + '_temp.txt'
                shell('mv {infile} {outfile}')
SnakeMake From line 129 of munge_stats/Snakefile 
140
141
run:
    format_list(input[0], output[0])
SnakeMake From line 140 of munge_stats/Snakefile 
151
152
153
154
run:
        for infile in input:
                outfile= params[0] + infile.split('-')[0].split('/')[-1] + '_temp.txt'
                shell('mv {infile} {outfile}')
SnakeMake From line 151 of munge_stats/Snakefile 
163
164
run:
    format_list(input[0], output[0])
SnakeMake From line 163 of munge_stats/Snakefile 
174
175
176
177
run:
        for infile in input:
                outfile= params[0] + infile.split('-')[0].split('/')[-1] + '_temp.txt'
                shell('mv {infile} {outfile}')
SnakeMake From line 174 of munge_stats/Snakefile 
186
187
188
189
190
191
192
193
194
195
196
197
198
199
run:
	d= pd.read_csv(input[0], header= 0, sep= '\t', usecols= ['#CHROM', 'POS', 'REF', 'ALT', 'AF_EXCLUDING_1000G'])
	d.columns= ['CHR', 'POS', 'REF', 'ALT', 'eaf']
	d['CHR']= np.where(d.CHR=='X', '23', d.CHR)
	KG= pd.read_csv(input[1], header= 0, sep='\t', compression= 'gzip', names= ['ID', 'ALT', 'REF', 'eaf'])
	KG['ID']= KG['ID'].str.replace(':ID', '')
	KG['ID']= KG['ID'].str.replace('X', '23')
	d['eaf']= np.where(d['REF']> d['ALT'], 1 - d.eaf, d.eaf)
	KG['eaf']= np.where(KG['REF']> KG['ALT'], 1- KG.eaf, KG.eaf)
	d['REF']= np.where(d.REF.str.len() > d.ALT.str.len(), 'I', d.REF)
	d['ALT']= np.where(d.REF.str.len()< d.ALT.str.len(), 'I', d.ALT)
	d['REF']= np.where(d.ALT== 'I', 'D', d.REF)
	d['ALT']= np.where(d.REF== 'I', 'D', d.ALT)
	KG['REF']= np.where(KG.REF.str.len() > KG.ALT.str.len(), 'I', KG.REF)
SnakeMake From line 186 of munge_stats/Snakefile 
229
230
script:
        'filter_SNPs.R'
SnakeMake From line 229 of munge_stats/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
library(data.table)
library(dplyr)

d= fread(snakemake@input[[1]])

x1= nrow(d)

d= arrange(d, CHR, POS, EFF, REF)

hrc= fread(snakemake@input[[2]], header=T)

d= inner_join(d, hrc, by= 'ID')
rm(hrc)
d$EAF= ifelse(is.na(d$EAF), d$eaf, d$EAF)

d[d$REF>d$EFF, c("REF", "EFF")]= d[d$REF > d$EFF, c("EFF", "REF")]

d$MAF= ifelse(d$EAF>0.5, 1- d$EAF, d$EAF)

d= filter(d, MAF>0.005)

d= filter(d, (MAF * 2 * N) > 6)

d$maf= ifelse(d$eaf> 0.5, 1 - d$eaf, d$eaf)
d$P= as.numeric(d$P)

d= filter(d, P<1, P>0)
d= filter(d, abs(MAF - maf) < 0.2)

d= select(d, -c(maf, MAF, eaf))

x2= nrow(d)

write.table(d, snakemake@output[[1]], col.names= T, row.names=F, sep= '\t', quote= F)

cohort= unlist(strsplit(unlist(strsplit(snakemake@input[[1]], '/'))[[10]], '_'))[2]
cat(c(cohort, '\t', x1, '\t', x2, '\n'), file= snakemake@output[[2]])
R dplyr data.table From line 1 of nonadditive/filter_SNPs.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
import pandas as pd
import numpy as np
import re

#d= pd.read_csv(snakemake.input[0], sep= '\t', header= 0)

#d['Allele1']= d['Allele1'].str.upper()
#d['Allele2']= d['Allele2'].str.upper()
#d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
#d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
#d['CHR']= d['CHR'].astype(str).astype(int)
#d['POS']= d['POS'].astype(str).astype(int)
#d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
#d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
#d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
#d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)

#d['CHR']= d['CHR'].astype(str).astype(int)
#d['POS']= d['POS'].astype(str).astype(int)

#d['pvalue']= d['pvalue'].astype(str).astype(float)

#d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
#d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF

#d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]

col_list= ['IMPACT', 'DISTANCE', 'SYMBOL', 'SYMBOL_SOURCE', 'BIOTYPE']
df_list= list()

for vep in pd.read_csv(snakemake.input[1], sep= '\t', header= None, names= ['Variation', 'Location', 'Allele', 'Gene', 'Feature', 'Feature_type', 'Consequence', 'cDNA_position', 'CDS_position', 'Protein_position', 'Amino_acids', 'Codons', 'Existing_variation', 'Extra'], comment= '#', chunksize= 100000):
	for i in col_list:
		vep[i]= vep['Extra'].apply(lambda y: dict([(x.split('=', 1)) for x in re.split(';(?=\w)', y) if x.find('=') > -1])[i] if i in y else '')
	vep= vep[['Variation', 'Location', 'Existing_variation', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'SYMBOL_SOURCE', 'BIOTYPE']]
	vep.columns= ['ID', 'Location', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'SYMBOL_SOURCE', 'BIOTYPE']
	vep['BIOTYPE1']= np.where(vep.BIOTYPE== 'protein_coding', 0, np.where(vep.BIOTYPE.str.contains('pseudo'), 2, 1))
	vep['DISTANCE']= np.where(vep.DISTANCE== '', 0, vep.DISTANCE)
	vep[['chr', 'pos', 'All']]= vep.ID.str.split('_', expand= True)
	vep[['EFF', 'REF']]= vep.All.str.split('/', expand= True)
	vep.loc[vep.REF > vep.EFF, ['REF', 'EFF']] = vep.loc[vep.REF > vep.EFF, ['EFF', 'REF']].values
	vep[['CHR', 'POS']]= vep['Location'].str.split(':', expand= True)
	vep['CHR']= np.where(vep['CHR']== 'X', '23', vep['CHR'])
	vep['ID']= vep.CHR.astype(int).astype(str) + ':' + vep.POS.astype(int).astype(str) + ':' + vep.REF + ':' + vep.EFF
	vep= vep[['ID', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'BIOTYPE', 'BIOTYPE1']]
	vep.sort_values(by= ['BIOTYPE1'], ascending= True, inplace= True)
	vep.drop_duplicates(subset= ['ID'], keep= 'first', inplace= True)
	df_list.append(vep)

vep= pd.concat(df_list)

vep.sort_values(by= ['BIOTYPE1'], ascending= True, inplace= True)
vep.drop_duplicates(subset= ['ID'], keep= 'first', inplace= True)
vep= vep[['ID', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'BIOTYPE']]


d= pd.read_csv(snakemake.input[0], sep= '\t', header= 0)
d['Allele1']= d['Allele1'].str.upper()
d['Allele2']= d['Allele2'].str.upper()
d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
d= d.loc[d.TOTALSAMPLESIZE> 66106, :]
d[['CHR', 'POS', 'REF','EFF']]= d['MarkerName'].str.split(':', expand= True)
d['CHR']= d['CHR'].astype(str).astype(int)
d['POS']= d['POS'].astype(str).astype(int)
d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'P-value']]
d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'pvalue']
d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
d['CHR']= d['CHR'].astype(str).astype(int)
d['POS']= d['POS'].astype(str).astype(int)
d['pvalue']= d['pvalue'].astype(str).astype(float)
d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
d['MAF']= np.where(d.EAF>0.5, 1 - d.EAF, d.EAF)
d= d.loc[d.MAF>= 0.1, :]
d= pd.merge(d, vep, on= ['ID'], how= 'left')
d.to_csv(snakemake.output[0], header=True, index= False, sep= '\t')
Python Pandas numpy Variant Effect Predictor (VEP) From line 1 of nonadditive/format_VEP.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	dcols= d.columns.values[1:]
	d.drop('INFO', 1, inplace= True)
	d.columns= dcols
	d= d.loc[d.INFO>= 0.4, :]
	d['MAF']=  np.where(d.EAF_CONTR> 0.5, 1- d.EAF_CONTR, d.EAF_CONTR)
	d= d.loc[d.MAF * 2 * d.N >6, :]
	d.drop('MAF', 1, inplace= True)
	d['REF_ALLELE']= np.where(d.REF_ALLELE.str.len()> d.EFF_ALLELE.str.len(), 'I', d.REF_ALLELE)
	d['EFF_ALLELE']= np.where(d.REF_ALLELE.str.len()< d.EFF_ALLELE.str.len(), 'I', d.EFF_ALLELE)
	d['REF_ALLELE']= np.where(d.EFF_ALLELE== 'I', 'D', d.REF_ALLELE)
	d['EFF_ALLELE']= np.where(d.REF_ALLELE== 'I', 'D', d.EFF_ALLELE)
	d['CHR']= d.CHR.apply(str)
	d['CHR']= np.where(d.CHR== 'X', '23', d.CHR)
	d['ID']= np.where(d.REF_ALLELE> d.EFF_ALLELE, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.EFF_ALLELE + ':' + d.REF_ALLELE, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.REF_ALLELE + ':' + d.EFF_ALLELE)
	d= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_DOM', 'P_VAL_REC', 'INFO']]
	df= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_REC', 'INFO']]
	d= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_DOM','INFO']]
	d= d.loc[(d.P_VAL_DOM!= '.' ), :]
	df= df.loc[(df.P_VAL_REC!= '.'), :]
	d[['BETA_ADD', 'P_VAL_DOM']]= d[['BETA_ADD', 'P_VAL_DOM']].apply(pd.to_numeric, errors= 'coerce')
	df[['BETA_ADD', 'P_VAL_REC']]= df[['BETA_ADD', 'P_VAL_REC']].apply(pd.to_numeric, errors= 'coerce')
	d.dropna(axis= 0, inplace= True)
	df.dropna(axis= 0, inplace= True)
	d.columns= ['ID', 'CHR', 'POS', 'EFF', 'REF', 'N', 'EAF', 'BETA', 'P', 'INFO']
	d.to_csv(output[0], sep= '\t', header= True, index= False)
	df.columns= ['ID', 'CHR', 'POS', 'EFF', 'REF', 'N', 'EAF', 'BETA', 'P', 'INFO']
	df.to_csv(output[1], sep= '\t', header= True, index= False)
SnakeMake From line 13 of nonadditive/Snakefile 
56
57
58
59
60
run:
	for i in range(len(input)):
		d= pd.read_csv(input[i], sep= '\t', header= 0)
		d= d.loc[d.INFO>= 0.4, :]
		d['MAF']=  np.where(d.EAF> 0.5, 1- d.EAF, d.EAF)
SnakeMake From line 56 of nonadditive/Snakefile 
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
run:
	for i in range(len(input)):
		print(input[i])
		d= pd.read_csv(input[i], header= 0, delim_whitespace= True)
		d[['BETA_ADD', 'P_VAL_DOM', 'P_VAL_REC', 'INFO', 'EAF_CONTR']]= d[['BETA_ADD', 'P_VAL_DOM', 'P_VAL_REC', 'INFO', 'EAF_CONTR']].apply(pd.to_numeric, errors= 'coerce')
		d= d.loc[d.INFO>= 0.4, :]
		d['MAF']=  np.where(d.EAF_CONTR> 0.5, 1- d.EAF_CONTR, d.EAF_CONTR)
		d= d.loc[d.MAF * 2 * d.N >6, :]
		d.drop('MAF', 1, inplace= True)
		d['REF_ALLELE']= np.where(len(d.REF_ALLELE)> len(d.EFF_ALLELE), 'I', d.REF_ALLELE)
		d['EFF_ALLELE']= np.where(len(d.REF_ALLELE)< len(d.EFF_ALLELE), 'I', d.EFF_ALLELE)
		d['REF_ALLELE']= np.where(d.EFF_ALLELE== 'I', 'D', d.REF_ALLELE)
		d['EFF_ALLELE']= np.where(d.REF_ALLELE== 'I', 'D', d.EFF_ALLELE)
		d['CHR']= d.CHR.apply(str)
		d['CHR']= np.where(d.CHR== '0X', 'X', d.CHR)
		d['CHR']= np.where(d.CHR== 'X', '23', d.CHR)
		d['ID']= np.where(d.REF_ALLELE> d.EFF_ALLELE, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.EFF_ALLELE + ':' + d.REF_ALLELE, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.REF_ALLELE + ':' + d.EFF_ALLELE)
		d= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_DOM', 'P_VAL_REC', 'INFO']]
		df= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_REC', 'INFO']]
		d= d[['ID', 'CHR', 'POS', 'EFF_ALLELE', 'REF_ALLELE', 'N', 'EAF_CONTR', 'BETA_ADD', 'P_VAL_DOM','INFO']]
		d= d.loc[(d.P_VAL_DOM!= '.' ), :]
		df= df.loc[(df.P_VAL_REC!= '.'), :]
		d.dropna(axis= 0, inplace= True)
		df.dropna(axis= 0, inplace= True)
		d.columns= ['ID', 'CHR', 'POS', 'EFF', 'REF', 'N', 'EAF', 'BETA', 'P', 'INFO']
		d.to_csv(output[i], sep= '\t', header= True, index= False)
		df.columns= ['ID', 'CHR', 'POS', 'EFF', 'REF', 'N', 'EAF', 'BETA', 'P', 'INFO']
		out= output[i].replace('dom', 'rec')
		df.to_csv(out, sep= '\t', header= True, index= False)
SnakeMake From line 105 of nonadditive/Snakefile 
145
146
script:
	'filter_SNPs.R'
SnakeMake From line 145 of nonadditive/Snakefile 
154
155
shell:
	'cat {input} > {output[0]}'
SnakeMake From line 154 of nonadditive/Snakefile 
165
166
167
168
shell:
    '''
    /home/pol/software/generic-metal/metal {input[0]} >> {output[1]}
    '''
SnakeMake From line 165 of nonadditive/Snakefile 
176
177
178
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0)
        d[['CHR', 'POS', 'REF', 'EFF']]= d['MarkerName'].str.split(':', expand= True)
SnakeMake From line 176 of nonadditive/Snakefile 
194
195
shell:
        '/home/pol/software/ensembl-vep/vep -i {input[0]} --check_existing --symbol --biotype --cache -O {output[0]} --offline --force_overwrite'
SnakeMake Variant Effect Predictor (VEP) From line 194 of nonadditive/Snakefile 
205
206
script:
	'format_VEP.py'
SnakeMake From line 205 of nonadditive/Snakefile 
214
215
216
217
218
219
220
221
222
223
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['MarkerName', 'Allele1'])
        d['CHR']= d.MarkerName.str.split(':').str[0]
        d['end']= d.MarkerName.str.split(':').str[1]
        d['CHR']= d.CHR.astype('str').astype('int')
        d['end']= d.end.astype('str').astype('int')
        d['start']= d.end - 1
        d.sort_values(by= ['CHR', 'start'], inplace= True)
        d= d[['CHR', 'start', 'end', 'MarkerName']]
        d.to_csv(output[0], sep= '\t', header= False, index= False)
SnakeMake From line 214 of nonadditive/Snakefile 
232
233
shell:
        'bedtools closest -t all -a {input[0]} -b {input[1]} > {output[0]}'
SnakeMake BEDTools From line 232 of nonadditive/Snakefile 
244
245
246
247
248
249
250
251
252
253
254
255
run:
        d= pd.read_csv(input[0], sep= '\t', header=0)
        rs= pd.read_csv(input[1], sep= '\t', header=0)
        d= pd.merge(d, rs, on= 'ID', how= 'left')
        d['RSID']= np.where(pd.isnull(d.RSID), d.name, d.RSID)
        d['RSID']= np.where(d.RSID== '', d.name, d.RSID)
        d['RSID']= np.where(d.RSID== '-', d.name, d.RSID)
        d.drop('name', 1, inplace= True)
        ne= pd.read_csv(input[2], sep= '\t', header= None, names= ['CHR', 'X', 'POS', 'ID', 'c1', 'p1', 'p2', 'nearestGene', 'Ensembl_gene'])
        ne= ne[['ID', 'nearestGene']]
        d= pd.merge(d, ne, on= 'ID', how= 'left')
        d.to_csv(output[0], sep= '\t', header= True, index= False, compression= 'gzip')
SnakeMake From line 244 of nonadditive/Snakefile 
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
run:
	for nfile in range(len(input)):
		d= pd.read_csv(input[nfile], sep= '\t', header= 0)
		d['Allele1']= d['Allele1'].str.upper()
		d['Allele2']= d['Allele2'].str.upper()
		d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
		d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
		d['CHR']= d['CHR'].astype(str).astype(int)
		d['POS']= d['POS'].astype(str).astype(int)
		d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
		d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
		d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
		d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
		d['CHR']= d['CHR'].astype(str).astype(int)
		d['POS']= d['POS'].astype(str).astype(int)
		d['pvalue']= d['pvalue'].astype(str).astype(float)
		d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
		d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
		d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
		d.to_csv(output[nfile], header=True, index= False, sep= '\t')
SnakeMake From line 14 of other_metas/Snakefile 
48
49
run:
	for fnumber in range(len(input)):
SnakeMake From line 48 of other_metas/Snakefile 
72
73
74
75
76
77
run:
	meta_files= [x for x in input if 'other_meta' in x]
	for nfile in range(len(meta_files)):
		meta= meta_files[nfile]
		out= output[nfile]
		shell('bedtools closest -t all -a {meta} -b {input[0]} > {out}')
SnakeMake BEDTools From line 72 of other_metas/Snakefile 
96
97
run:
	rs= pd.read_csv(input[0], sep= '\t', header=0)
SnakeMake From line 96 of other_metas/Snakefile 
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
run:
        for nfile in range(len(input)):
                d= pd.read_csv(input[nfile], sep= '\t', header= 0)
                d['Allele1']= d['Allele1'].str.upper()
                d['Allele2']= d['Allele2'].str.upper()
                d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
                d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
                d['CHR']= d['CHR'].astype(str).astype(int)
                d['POS']= d['POS'].astype(str).astype(int)
                d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
                d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
                d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
                d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
                d['CHR']= d['CHR'].astype(str).astype(int)
                d['POS']= d['POS'].astype(str).astype(int)
                d['pvalue']= d['pvalue'].astype(str).astype(float)
                d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
                d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
                d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
                d.to_csv(output[nfile], header=True, index= False, sep= '\t')
SnakeMake From line 116 of other_metas/Snakefile 
144
145
146
147
148
149
150
run:
        rs= pd.read_csv(input[0], sep= '\t', header=0)
        d= pd.read_csv(input[1], sep= '\t', header=0)
        d= pd.merge(d, rs, on= 'ID', how= 'left')
        d['RSID']= d.name
        d.drop('name', 1, inplace= True)
        d.to_csv(output[0], sep= '\t', header= True, index= False, compression= 'gzip')
SnakeMake From line 144 of other_metas/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
library(data.table)
library(dplyr)
library(coloc)
library(parallel)

df= fread(snakemake@input[[1]], select= c('CHR', 'POS', 'ID', 'BETA', 'SE', 'TOTALSAMPLESIZE', 'EAF'))

df= filter(df, !duplicated(ID))

df$MAF= ifelse(df$EAF>0.5, 1 - df$EAF, df$EAF)

x= fread(snakemake@input[[2]], select= c('CHR', 'POS', 'nearestGene'))
x= x[, c('CHR', 'POS', 'nearestGene')]
names(x)= c('CHR', 'pos2', 'nearestGene')

df= inner_join(df, x, by= 'CHR')

df= filter(df, POS>= pos2 - 1.5*10**6, POS< pos2 + 1.5*10**6)

z= fread(snakemake@input[[3]], select= c('chr', 'pos', 'Allele1', 'Allele2', 'Freq1', 'Effect', 'StdErr', 'TotalSampleSize'))

z$Allele1= toupper(z$Allele1)
z$Allele2= toupper(z$Allele2)

z$ID= with(z, ifelse(Allele1 > Allele2, paste(chr, pos, Allele2, Allele1, sep= ':'), paste(chr, pos, Allele1, Allele2, sep= ':')))

z$maf= ifelse(z$Freq1> 0.5, 1 - z$Freq1, z$Freq1)

z= select(z, ID, maf, Effect, StdErr, TotalSampleSize)

df= inner_join(df, z, by= 'ID')

rm(z)

pph_outfile= snakemake@output[[1]]
results_outfile= snakemake@output[[2]]


cat('nsnps\tPP.H0.abf\tPP.H1.abf\tPP.H2.abf\tPP.H3.abf\tPP.H4.abf\tprotein\n', file = snakemake@output[[1]])

cat('snp\tV.df\tz.df1\tr.df1\tlABF.df1\tV.df2\tz.df2\tr.df2\tlABF.df2\tinternal.sum.lABF\tSNP.PP.H4\tprotein\n', file= snakemake@output[[2]])

prior1= 1 * 10**-4
prior2= 1 * 10**-4
prior12= 5 * 10**-6


df= data.frame(df)


colocalization_eqtl= function(temp_df){
	protein= unique(temp_df$nearestGene)
        if (nrow(temp_df)== 0) {

        PPH= data.frame(nsnps= 0, PP.H0.abf= 0,PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')

        } else {
	temp_df = filter(temp_df, SE>0, StdErr> 0)

	if (grepl('allPTD', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, s= 0.067, MAF= temp_df$MAF)
        } else if (grepl('postTerm', snakemake@input[[1]])) {
        data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N=temp_df$TOTALSAMPLESIZE, type= 'cc', snp= temp_df$ID, s= 0.122, MAF= temp_df$MAF)
        } else {data1= list(beta= temp_df$BETA, varbeta= temp_df$SE**2, N= temp_df$TOTALSAMPLESIZE, type= 'quant', snp= temp_df$ID, MAF= temp_df$MAF) }

        data2= list(beta= temp_df$Effect, varbeta= temp_df$StdErr**2, N=temp_df$TotalSampleSize, type= 'quant', snp= temp_df$ID, MAF= temp_df$maf)
        myres= tryCatch({suppressWarnings(coloc.abf(data1, data2, p1= prior1, p2= prior2, p12= prior12))}, error= function(e) { return(0)}
)
        if (length(myres)==1 ) { 
        PPH= data.frame(nsnps= 0, PP.H0.abf= 0, PP.H1.abf= 0, PP.H2.abf= 0, PP.H3.abf= 0, PP.H4.abf= 0, protein= protein)
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= data.frame(snp= 'none', V.df1= 0, z.df1= 0, r.df1= 0, lABF.df1= 0, V.df2= 0, z.df2= 0, r.df2= 0, lABF.df2= 0, internal.sum.lABF= 0, SNP.PP.H4= 0, protein= protein)
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        print('next')
        } else {
        PPH= data.frame(t(myres[[1]]))
        PPH$protein= protein
        fwrite(PPH, pph_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
        res= myres[[2]]
        res$protein= protein
        fwrite(res, results_outfile, sep= '\t', row.names=F, col.names= F, quote=F, append= T)
}
}
}



mclapply(split(df, df$nearestGene), colocalization_eqtl, mc.cores= 3)
R dplyr data.table Quant coloc From line 1 of pQTLs/coloc_pQTL.R 
12
13
script:
	'coloc_pQTL.R'
SnakeMake From line 12 of pQTLs/Snakefile 
18
19
20
21
22
23
24
body .main-container {
  max-width: 1280px !important;
  width: 1280px !important;
}
body {
  max-width: 1280px !important;
}
R Markdown From line 18 of reports/all_files_QC.Rmd 
28
pheno= unlist(strsplit(snakemake@input[[2]], '/'))[8]
R Markdown From line 28 of reports/all_files_QC.Rmd 
40
41
42
43
44
45
46
47
48
49
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library("data.table")
library(moments)
options(warn=-1)
opts_chunk$set(fig.width = 12)
R Markdown ggplot2 dplyr data.table tidyr cowplot knitr kableExtra From line 40 of reports/all_files_QC.Rmd 
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
p1_list= list()
p2_list= list()
p3_list= list()
df_list= list()
df2_list= list()
fl= length(snakemake@input)

dec= fread(snakemake@input[[grep('DECODE', snakemake@input)]])
dec$MAF= ifelse(dec$EAF>0.5, 1 - dec$EAF, dec$EAF)
dec= select(dec, c(SNP, BETA))
names(dec)= c('SNP', 'BETA_dec')
for (i in 2:fl){

df= fread(snakemake@input[[i]])
df= select(df, -c(STRAND))

cohort= unlist(strsplit(snakemake@input[[i]], '/'))[9]

df$cohort= cohort
df= filter(df, !is.na(EAF))

p1= summarize(df, n_m= median(N, na.rm=T), se_m= mean(SE, na.rm=T))
p1$cohort= cohort
p2= summarize(df, N_max= sqrt(max(N)), EAF_m= median(1/sqrt(2*EAF*(1-EAF)), na.rm=T) / median(SE))
p2$cohort= cohort

d= filter(df, pvalue> median(pvalue, na.rm=T))

p3= summarize(d,SK= skewness(BETA/SE), KU= kurtosis(BETA/SE))
p3$cohort= cohort

p1_list[[cohort]]= p1
p2_list[[cohort]]= p2
p3_list[[cohort]]= p3

df_list[[cohort]]= group_by(df, CHR) %>% sample_n(5000, replace=T)

df= inner_join(df, dec, by= 'SNP')
df$beta_diff= df$BETA - df$BETA_dec

df2_list[[cohort]]= group_by(df, CHR) %>% sample_n(5000, replace=T)

}


p1= do.call("rbind", p1_list)
p2= do.call("rbind", p2_list)
p3= do.call("rbind", p3_list)
d= do.call("rbind", df_list)
d2= do.call('rbind', df2_list)
R Markdown From line 57 of reports/all_files_QC.Rmd 
113
114
115
116
117
118
ggplot(p1, aes(n_m, se_m)) +
geom_point() +
geom_text(aes(label=cohort), hjust=0, vjust=0) +
theme_cowplot() +
xlab('Median(N)') +
ylab('Mean(SE)')
R Markdown ggplot2 From line 113 of reports/all_files_QC.Rmd 
127
128
129
130
131
132
ggplot(p2, aes(N_max, EAF_m)) +
geom_point() +
geom_text(aes(label=cohort), hjust=0, vjust=0) +
theme_cowplot() +
xlab('SQRT(Max(N))') +
ylab('median(1/sqrt(2*EAF*(1-EAF)), na.rm=T) / median(SE)')
R Markdown ggplot2 From line 127 of reports/all_files_QC.Rmd 
145
146
147
148
149
150
ggplot(p3, aes(SK, KU)) +
geom_point() +
geom_text(aes(label=cohort), hjust=0, vjust=0) +
theme_cowplot() +
xlab('Skewness (Z-score)') +
ylab('Kurtosis (Z-score)')
R Markdown ggplot2 From line 145 of reports/all_files_QC.Rmd 
160
161
162
163
164
165
166
167
d$MAF= ifelse(d$EAF>0.5, 1 - d$EAF, d$EAF)

ggplot(d, aes(MAF, BETA)) +
geom_point() +
facet_wrap(vars(cohort), scales= 'free_y', ncol= 3) +
theme_cowplot() +
xlab('MAF') +
ylab('BETA')
R Markdown ggplot2 From line 160 of reports/all_files_QC.Rmd 
177
178
179
180
181
182
183
184
d2$MAF= ifelse(d2$EAF>0.5, 1 - d2$EAF, d2$EAF)

ggplot(d2, aes(MAF, beta_diff)) +
geom_point() +
facet_wrap(vars(cohort), scales= 'free_y', ncol= 3) +
theme_cowplot() +
xlab('MAF') +
ylab('BETA cohort - BETA DECODE')
R Markdown ggplot2 From line 177 of reports/all_files_QC.Rmd 
10
11
12
13
14
15
16
17
18
19
20
library(tint)

knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tint'))
options(htmltools.dir.version = FALSE)
pheno= snakemake@wildcards[['pheno']]
pdf.options(useDingbats = TRUE)
#knitr::opts_chunk$set(dpi=300)
knitr::opts_chunk$set(dev = 'png', warning= FALSE, message= FALSE, dpi= 600)


pheno= ifelse(pheno=='allPTD', 'Preterm Delivery', ifelse(pheno== 'postTerm', 'Post Term', ifelse(pheno=='GAraw', 'Gestational duration', 'Normalized Gestational Duration')))
R Markdown From line 10 of reports/coloc.Rmd 
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library(ggrepel)
library("data.table")
options(warn=-1)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

d= fread(snakemake@input[[1]])

d$p1= gsub('.txt.sumstats.gz', '', apply(d[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
d$p2= gsub('.txt.sumstats.gz', '', apply(d[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))

d$trait= d$p2
d$trait= with(d, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Birth weight maternal effect',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
		ifelse(trait== 'SHBG_male', 'SHBG (men)',
		ifelse(trait== 'CBAT_fem', 'CBAT (women)',
		ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
		ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
		ifelse(trait== 'BW_fetal', 'Birth weight fetal effect',
		ifelse(trait== 'BW_fetal_effect', 'Birth weight fetal effect (adjusted MG)',
		ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
		ifelse(trait== 'BW_maternal_effect', 'Birth weight maternal effect (adjusted FG)',
                ifelse(trait== 'PCOS', 'Polycistic ovary syndrome', 'Endometriosis'))))))))))))))))))))))
R Markdown ggplot2 dplyr data.table tidyr cowplot ggrepel knitr kableExtra From line 30 of reports/coloc.Rmd 
78
cat(paste0('\n- ', factor(d$trait)), sep= "\n")
R Markdown From line 78 of reports/coloc.Rmd 
82
Testosterone in males was further included as a negative control, and only after a first round of genetic correlations.
R Markdown From line 82 of reports/coloc.Rmd 
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
bw= filter(d, grepl('Birth weight ', d$trait))
df= filter(d, !grepl('Birth weight ', d$trait))

df$significant= ifelse(df$p< 0.05 / (nrow(df) - 7 ), '1', '0')

bw$trait= with(bw, ifelse(trait== 'Birth weight fetal effect', 'Fetal effect', ifelse(
			trait== 'Birth weight fetal effect (adjusted MG)', 'Fetal effect\n (adjusted MG)', ifelse(
			trait=='Birth weight maternal effect (adjusted FG)', 'Maternal effect \n(adjusted FG)', ifelse(
			trait== 'Birth weight maternal effect', 'Maternal effect', '')))))


bw$trait= gsub('Birth weight', '', bw$trait)
ggplot(bw, aes(trait, rg, colour= trait)) +
geom_point() +
geom_errorbar(aes(ymin= I(rg - 1.96*se) , ymax= (rg + 1.96 * se)), width=.1, position=position_dodge(.9)) +
theme_cowplot() +
scale_colour_manual(guide= F, values= colorBlindBlack8[c(1,2,4,8)]) +
xlab('Birth weight') +
ylab('Genetic correlation [95% CI]') +
geom_hline(yintercept= 0) +
ylim(-1, 1) +
scale_y_continuous(breaks= seq(-1, 1, 0.2)) +
geom_hline(yintercept= seq(-1, 1, 0.2), colour= 'grey', size= 0.3, linetype= 'dashed')
R Markdown ggplot2 From line 86 of reports/coloc.Rmd 
115
116
117
118
119
120
121
122
123
124
125
126
ggplot(df, aes(trait, rg, colour= significant)) +
geom_point() +
geom_errorbar(aes(ymin= I(rg - 1.96*se) , ymax= (rg + 1.96 * se)), width=.2, position=position_dodge(.9)) +
theme_cowplot() +
theme(axis.text.x= element_text(angle=45, vjust= 1, hjust= 1)) +
scale_colour_manual(guide= F, values= c('#737373', colorBlindBlack8[2])) +
xlab('Reproductive traits') +
ylab('Genetic correlation [95% CI]') +
geom_hline(yintercept= 0) +
ylim(-1, 1) +
scale_y_continuous(breaks= seq(-1, 1, 0.2)) +
geom_hline(yintercept= seq(-1, 1, 0.2), colour= 'grey', size= 0.3, linetype= 'dashed')
R Markdown ggplot2 From line 115 of reports/coloc.Rmd 
132
We used a Bonferroni corrected threshold for significance (0.05/ 13). We exclude testosterone in males, as this test was performed a posteriori as a negative control for testosterone in women.
R Markdown From line 132 of reports/coloc.Rmd 
157
While coloc naiveley assumes one causal variant, it does not require an LD matrix that represents the summary statistics used. This is almost impossible to obtain without an LD matrix from each of the studies used in the meta-analysis.
R Markdown From line 157 of reports/coloc.Rmd 
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
inputs= snakemake@input[grep('pph', snakemake@input)]

df_list= list()

for (infile in inputs) {
d_temp= fread(infile)
#d_temp$trait= gsub('.txt', '', unlist(strsplit(infile, '_'))[2])
df_list[[infile]]= d_temp
}

d= do.call('rbind', df_list)

d$trait= with(d, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Birth weight maternal effect',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
		ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
		ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
		ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Birth weight fetal effect',
                ifelse(trait== 'BW_fetal_effect', 'Birth weight fetal effect (adjusted MG)',
                ifelse(trait== 'BW_maternal_effect', 'Birth weight maternal effect (adjusted FG)',
                ifelse(trait== 'PCOS', 'Polycistic ovary syndrome', 'Endometriosis'))))))))))))))))))))))

d$locus= gsub('_', ' ', d$locus)
d$locus= gsub('chr', 'Chr', d$locus)
d$locus= gsub('23', 'X', d$locus)
R Markdown From line 162 of reports/coloc.Rmd 
203
We identify the different loci as the chromosome where the locus is located and the nearest protein coding gene to the top associated genetic variant (e.g., Chr5 EBF1).
R Markdown From line 203 of reports/coloc.Rmd 
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
x1= select(d, locus, PP.H3.abf, trait)
x1$PP= 'Shared locus'

x2= select(d, locus, PP.H4.abf, trait)
x2$PP= 'Shared genetic variant'

names(x1)= c('locus', 'coloc', 'trait', 'PP')
names(x2)= c('locus', 'coloc', 'trait', 'PP')

x= bind_rows(x1, x2)

d$coloc= with(d, ifelse(PP.H4.abf< 0.5, 'No evidence', ifelse(PP.H4.abf>=  0.50 & PP.H4.abf< 0.75, 'Suggestive evidence', 'Strong evidence')))

ggplot(d, aes(trait, locus, size= PP.H4.abf,  fill= coloc, color= coloc, shape= direction, alpha= coloc)) + 
geom_point() + 
theme_cowplot() + 
theme(axis.text.x= element_text(angle=45, vjust= 1, hjust= 1)) +
scale_size_binned('Posterior probability of colocalization', guide= FALSE) +
scale_alpha_manual('Colocalization', values= c(1,0.55, 0.55)) +
scale_shape_manual('Effect direction', values=c(25, 21, 24)) +
scale_fill_manual('Colocalization', values=  c('#737373', colorBlindBlack8[2], colorBlindBlack8[4])) +
scale_colour_manual('Colocalization', values=  c('#737373', colorBlindBlack8[2], colorBlindBlack8[4])) +
xlab('') +
ylab('')
R Markdown ggplot2 From line 208 of reports/coloc.Rmd 
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
x3= select(d, locus, PP.H3.abf, trait)
x3$PP= 'H3'

x4= select(d, locus, PP.H4.abf, trait)
x4$PP= 'H4'

x0= select(d, locus, PP.H0.abf, trait)
x0$PP= 'H0'

x1= select(d, locus, PP.H1.abf, trait)
x1$PP= 'H1'

x2= select(d, locus, PP.H2.abf, trait)
x2$PP= 'H2'

names(x0)= c('locus', 'coloc', 'trait', 'PP')
names(x1)= c('locus', 'coloc', 'trait', 'PP')
names(x2)= c('locus', 'coloc', 'trait', 'PP')
names(x3)= c('locus', 'coloc', 'trait', 'PP')
names(x4)= c('locus', 'coloc', 'trait', 'PP')

x= bind_rows(x0, x1, x2, x3, x4)


x= x[order(x$PP, decreasing= T),]

x$evidence= ifelse(x$coloc>= 0.75, '1', '0')

ggplot(filter(x, PP== 'H3' | PP== 'H4'), aes(fill= factor(PP), y=coloc, x= locus, alpha= evidence)) + 
geom_bar(position="stack", stat="identity") +
scale_fill_manual('Posterior probability', values= c(colorBlindBlack8[2], colorBlindBlack8[4])) +
scale_alpha_manual('Posterior probability', values= c(0.55, 0.8), guide=FALSE) +
facet_wrap(vars(trait), ncol= 3) +
theme_cowplot() +
theme(axis.text.x= element_text(angle=45, vjust= 1, hjust= 1),
strip.background= element_blank(),
legend.position= 'bottom') +
xlab('') +
ylab('')
R Markdown ggplot2 From line 248 of reports/coloc.Rmd 
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
z= filter(d, PP.H4.abf >= 0.75)


res_inputs= snakemake@input[grep('results_', snakemake@input)]

df_list= list()

for (infile in res_inputs) {
x= fread(infile, select= c('snp', 'z.df1', 'z.df2', 'SNP.PP.H4', 'locus', 'trait'))

x= gather(x, pheno, zscore, c(z.df1, z.df2))
x$pvalue=2 *  pnorm(-abs(x$zscore))
x= separate(x, snp, into= c('CHR', 'POS', 'REF', 'EFF'), sep =':', remove= F)
x$POS= as.numeric(x$POS)
x$pheno= ifelse(x$pheno =='z.df1', pheno, x$trait)
df_list[[infile]]= x
}

df= bind_rows(df_list)

df$trait= with(df, ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Birth weight maternal effect',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
		ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
		ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'BW_fetal', 'Birth weight fetal effect',
                ifelse(trait== 'BW_fetal_effect', 'Birth weight fetal effect (adjusted MG)',
		ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_maternal_effect', 'Birth weight maternal effect (adjusted FG)',
                ifelse(trait== 'PCOS', 'Polycistic ovary syndrome', 'Endometriosis'))))))))))))))))))))))

df$locus= gsub('_', ' ', df$locus)
df$locus= gsub('chr', 'Chr', df$locus)
df$locus= gsub('23', 'X', df$locus)

z$id= paste(z$trait, z$locus, sep= ':')
df$id= paste(df$trait, df$locus, sep= ':')
df= filter(df, id %in% z$id)

df$pheno= ifelse(df$pheno== pheno, pheno, df$trait)

for (i in unique(df$id)){


PP= filter(z, id== i)$PP.H4.abf

temp_df= filter(df, id== i)
temp_df$pheno= temp_df$pheno
temp_df$POS= temp_df$POS / 10**6
high_df= filter(temp_df, id == i, SNP.PP.H4== max(SNP.PP.H4))

(ggplot() + 
geom_point(data= temp_df, aes(POS, -log10(pvalue), colour= pheno), size= 1, alpha = 0.5) +
geom_point(data= high_df, aes(POS, -log10(pvalue)), colour= colorBlindBlack8[1], size= 2) + 
facet_wrap(vars(pheno), nrow= 2, scales = "free_y") +
theme_cowplot(font_size=14) +
theme(strip.background= element_blank()) +
scale_colour_manual(guide=FALSE, values= colorBlindBlack8[c(4,2)]) +
ylab('-log10(pvalue)') +
xlab('Position (Mbp)') +
geom_text_repel(data=high_df, aes(x= POS, y= -log10(pvalue), label=snp), hjust= 0.5, size=3, vjust= 1) +
ggtitle(paste('Locus: ', unique(temp_df$locus), '. Posterior probability for shared causal variant: ', round(PP, 3)))) %>% print()


cat('  \n')


}
R Markdown From line 297 of reports/coloc.Rmd 
18
19
cohort= unlist(strsplit(snakemake@input[[1]], '/'))[9]
pheno= unlist(strsplit(snakemake@input[[1]], '/'))[8]
R Markdown From line 18 of reports/file_level_qc.Rmd 
29
30
31
32
33
34
35
36
37
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library("data.table")
options(warn=-1)
#opts_chunk$set(dpi=300, out.width="300px")
R Markdown ggplot2 dplyr data.table tidyr cowplot knitr kableExtra From line 29 of reports/file_level_qc.Rmd 
45
d= fread(snakemake@input[[1]], h=T)
R Markdown From line 45 of reports/file_level_qc.Rmd 
53
kable(summary(select(d, BETA, SE, pvalue, EAF, N)))
R Markdown From line 53 of reports/file_level_qc.Rmd 
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
#dec= fread(snakemake@input[[3]],h=T, select= c('CHR', 'POS', 'BETA', 'SE', 'EFF', 'REF', 'EAF', 'pvalue'))

#names(dec)= c('CHR', 'POS', 'BETA_dec', 'SE_dec', 'A1_dec', 'A2_dec', 'EAF_dec', 'pvalue_dec')

#dec$BETA_dec= ifelse(dec$A1_dec> dec$A2_dec, dec$BETA_dec, -1* dec$BETA_dec)
#dec$EAF_dec= ifelse(dec$A1_dec> dec$A2_dec, dec$EAF_dec, 1- dec$EAF_dec)

x= fread(snakemake@input[[2]], h=T)


df= d
df$CHR= as.numeric(df$CHR)
x$CHR= as.numeric(x$CHR)
df$EAF= ifelse(df$EFF> df$REF, df$EAF, 1 - df$EAF)

df= inner_join(df, x, on= c('CHR', 'POS'))

df= filter(df, (EFF== ea & REF== oa) | (REF== ea & EFF== oa))

if (nrow(df)>0){
ggplot(df, aes(eaf, EAF)) +
geom_point(alpha= 1/10) +
theme_cowplot(12) +
xlab('EAF HRC') +
ylab('EAF Sample')
} else{
print('No match')
}

rm(df)
rm(x)
R Markdown ggplot2 From line 63 of reports/file_level_qc.Rmd 
101
102
103
104
105
106
107
108
109
110
111
112
113
114
d$BETA= ifelse(d$REF> d$EFF, d$BETA, -1* d$BETA)
#dec= inner_join(d, dec, on= c('CHR', 'POS'))

#dec= filter(dec, EFF== A1_dec, REF== A2_dec, pvalue_dec< 0.001)

#dec= group_by(dec, CHR) %>% slice_sample(n= 1000, replace= T)

#ggplot(dec, aes(BETA_dec/SE_dec, BETA/SE)) +
#geom_point(alpha= 1/10) +
#theme_cowplot(12) +
#xlab('DECODE z-score') +
#ylab('Observed z-score')

#rm(dec)
R Markdown From line 101 of reports/file_level_qc.Rmd 
125
126
127
128
129
130
131
d$exp_pvalue= pchisq((d$BETA/d$SE)^2, df=1, lower=F)

ggplot(group_by(d, CHR) %>% sample_n(10000, replace=T ), aes(-log10(exp_pvalue), -log10(pvalue))) +
geom_point(alpha= 1/10) +
theme_cowplot() +
xlab('Expected pvalue') +
ylab('Observed pvalue')
R Markdown ggplot2 From line 125 of reports/file_level_qc.Rmd 
140
141
142
143
144
145
146
147
df= arrange(d, pvalue) %>% mutate(exp1= -log10(1:length (pvalue)/length (pvalue)))

ggplot(filter(df, pvalue<0.05), aes(exp1, -log10(pvalue))) +
  geom_point(size= 0.4) +
  geom_abline(intercept = 0, slope = 1, alpha = .5) +
theme_cowplot(12, font_size= 12) +
xlab('Expected (-log10(p-value))') +
ylab('Observed (-log10(p-value))')
R Markdown ggplot2 From line 140 of reports/file_level_qc.Rmd 
10
11
12
13
14
15
16
17
18
19
20
library(tint)

knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tint'))
options(htmltools.dir.version = FALSE)
pheno= snakemake@wildcards[['pheno']]
pdf.options(useDingbats = TRUE)
#knitr::opts_chunk$set(dpi=300)
knitr::opts_chunk$set(dev = 'png', warning= FALSE, message= FALSE, dpi= 600)


pheno= ifelse(pheno=='allPTD', 'Preterm Delivery', ifelse(pheno== 'postTerm', 'Post Term', ifelse(pheno=='GAraw', 'Gestational duration', 'Normalized Gestational Duration')))
R Markdown From line 10 of reports/forest_plots.Rmd 
29
30
31
32
33
34
35
36
37
38
39
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library(ggrepel)
library("data.table")
options(warn=-1)

colorBlindBlack8  <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
R Markdown ggplot2 dplyr data.table tidyr cowplot ggrepel knitr kableExtra From line 29 of reports/forest_plots.Rmd 
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
d= fread(snakemake@input[[1]])

z= fread(snakemake@input[[3]])




df= fread(snakemake@input[[2]], select= (c('MarkerName', 'Effect', 'StdErr', 'HetISq', 'HetPVal', 'TOTALSAMPLESIZE', 'P-value', 'Allele1', 'Allele2')))
names(df)= c('SNP', 'BETA', 'SE', 'HetISq', 'HetPval', 'N', 'pvalue', 'A1', 'A2')
df= filter(df, SNP %in% d$SNP)

df= separate(df, SNP, into= c('CHR', 'POS', 'Ax1', 'Ax2', 'ID'), sep= ':', remove= F)
df$BETA= ifelse(df$A2 > df$A1, -1 * df$BETA, df$BETA)
df$CHR= ifelse(df$CHR== 'X','23', df$CHR)
df$CHR= as.integer(df$CHR)
df$POS= as.integer(df$POS)
df= select(df, -c(A1, A2, ID, Ax1, Ax2))

df$cohort= 'Meta-analysis'
d= bind_rows(d, df)

z$CHR= ifelse(z$CHR== 'X','23', z$CHR)
z$CHR= as.integer(z$CHR)

d= inner_join(d, z, by= 'CHR') %>% filter(POS> pos1, POS< pos2)



d$locus= paste0('Chr ', d$CHR,': ', d$nearestGene)
d$cohort= paste0(d$cohort, ' (n= ', d$N, ')')
R Markdown From line 44 of reports/forest_plots.Rmd 
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
for (i in unique(d$locus)) {
temp_df= d[d$locus== i, ]

cat('\n')

cat("\n# Forest plot for locus ", i, "\n")

cat("\n")

cat('\n')

cat('Lead variant: \n', temp_df[!is.na(temp_df$HetISq), ]$SNP)

cat('\n')

cat(paste0('\n Meta-analysis: Beta= ', round(temp_df[!is.na(temp_df$HetISq), ]$BETA, 3), ' (95% CI= ', round(temp_df[!is.na(temp_df$HetISq), ]$BETA - 1.96 * temp_df[!is.na(temp_df$HetISq), ]$SE, 3), ', ', round(temp_df[!is.na(temp_df$HetISq), ]$BETA + 1.96 * temp_df[!is.na(temp_df$HetISq), ]$SE, 3), '); pvalue= ', temp_df[!is.na(temp_df$HetISq), ]$pvalue))

cat('\n')

cat('\n')
temp_df= temp_df[order(temp_df$N, decreasing= T), ]

(ggplot(temp_df, aes(x=factor(cohort, level = factor(cohort)), y=BETA, ymin= BETA - 1.96 * SE, ymax= BETA + 1.96 * SE, colour= !is.na(HetISq), shape= !is.na(HetISq)), alpha= 0.5) +
 geom_pointrange(size= 1, alpha= 0.7) +
scale_shape_manual(values= c(15, 18), guide= F) +
 geom_hline(yintercept = 0, linetype=2) +
scale_y_continuous(sec.axis = dup_axis()) +
 coord_flip() +
scale_colour_manual(values= c(colorBlindBlack8[3], colorBlindBlack8[4]), guide= F) +
theme_cowplot() +
 xlab('') +
    ylab('Beta [95% CI]') +
geom_vline(xintercept= 0, linetype= "dotted", colour= 'grey') ) %>% print()

cat('\n')

cat('Test for heterogeneity: I^2^= ', temp_df[!is.na(temp_df$HetISq), ]$HetISq, '%; Het pvalue= ', temp_df[!is.na(temp_df$HetISq), ]$HetPval, '\n')

cat('\\pagebreak')

}
R Markdown From line 78 of reports/forest_plots.Rmd 
10
11
12
13
14
15
16
17
18
19
20
library(tint)

knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tint'))
options(htmltools.dir.version = FALSE)
pheno= snakemake@wildcards[['pheno']]
pdf.options(useDingbats = TRUE)
#knitr::opts_chunk$set(dpi=300)
knitr::opts_chunk$set(dev = 'png', warning= FALSE, message= FALSE, dpi= 600)


pheno= ifelse(pheno=='allPTD', 'Preterm Delivery', ifelse(pheno== 'postTerm', 'Post Term', ifelse(pheno=='GAraw', 'Gestational duration', 'Normalized Gestational Duration')))
R Markdown From line 10 of reports/meta_qc.Rmd 
32
33
34
35
36
37
38
39
40
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library(ggrepel)
library("data.table")
options(warn=-1)
R Markdown ggplot2 dplyr data.table tidyr cowplot ggrepel knitr kableExtra From line 32 of reports/meta_qc.Rmd 
45
46
47
48
49
50
51
52
53
54
55
d= fread(snakemake@input[[1]], h= T)
d$MAF= ifelse(d$EAF>0.5,  1 - d$EAF, d$EAF)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

ggplot(d, aes(MAF)) +
  geom_density(fill= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
xlab('MAF')

#d= select(d, -MAF)
R Markdown ggplot2 From line 45 of reports/meta_qc.Rmd 
61
62
63
64
ggplot(d, aes(TOTALSAMPLESIZE)) +
  geom_density(fill= colorBlindBlack8[3]) +
theme_cowplot(font_size= 12) +
xlab('Sample size')
R Markdown ggplot2 From line 61 of reports/meta_qc.Rmd 
72
73
74
75
ggplot(d, aes(BETA)) +
  geom_density(fill= colorBlindBlack8[2]) +
theme_cowplot(font_size= 12) +
xlab('Beta')
R Markdown ggplot2 From line 72 of reports/meta_qc.Rmd 
91
92
**Effective sample size** for binary phenotypes was calculated as:  
$$\frac{2}{(\frac{1}{Ncases} + \frac{1}{Ncontrols})}$$
R Markdown From line 91 of reports/meta_qc.Rmd 
110
111
112
ge= data.frame(CHR= c(5, 3, 1, 23, 1), pos_ge= c(157895049, 127881613, 22470407, 115164770, 22470407))

#kable(summary(select(d, BETA, SE, pvalue, EAF, TOTALSAMPLESIZE)), digits = c(3, 3, 5, 4, 0), col.names = c('Beta', 'Standard error', 'P-value', 'Effect allele frequency', 'Sample size'), caption= 'Summary statistics after QC.')
R Markdown From line 110 of reports/meta_qc.Rmd 
124
The same number of loci is obtained when using a larger radius (1.5Mb).
R Markdown From line 124 of reports/meta_qc.Rmd 
128
We note that we used a naive approach to identify independent loci. This should be interpreted cautiously. We mapped top genetic variants to the body (TSS or TES) nearest protein coding gene.
R Markdown From line 128 of reports/meta_qc.Rmd 
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
df= arrange(d, pvalue)

#df= group_by(df, CHR, POS) %>% filter(row_number() == 1) %>% ungroup()
df= df[!duplicated(df[, c('CHR', 'POS')]), ]

dg= df %>% arrange(CHR, POS) %>% filter(pvalue< 5*10**-8) %>% group_by(CHR) %>%
		mutate(d=POS-lag(POS, default=-Inf), clumpid=cumsum(d>250000)) %>%
		group_by(CHR, clumpid) %>%
		filter(rank(pvalue, ties.method = "random")==1) %>%
		mutate(GENE= ifelse(SYMBOL=='', RSID, SYMBOL)) %>%
		ungroup()

dg= filter(dg, pvalue< 5*10**-8)
dg= group_by(dg, CHR, POS) %>% filter(row_number()== 1)
#dg$GENE= ifelse(grepl('rs|-', dg$GENE), dg$nearestGene, ifelse(dg$GENE=='', dg$nearestGene, dg$GENE))
dg$GENE= dg$nearestGene

  don <- df %>%
    group_by(CHR)      %>%
    summarise(chr_len= max(POS)) %>%
    mutate(tot= cumsum(as.numeric(chr_len))-chr_len) %>% # Calculate cumulative position of each chromosome
    select(-chr_len) %>%
    left_join(df, ., by= 'CHR') %>%
    arrange(CHR, POS) %>% # Add a cumulative position of each SNP
    mutate( BPcum=POS+tot) %>%
	 ungroup()

  axisdf = don %>% group_by(CHR) %>% summarize(center=( max(BPcum) + min(BPcum) ) / 2 )
  names(axisdf)= c('CHR', 'center')
HC= -log10(5*10**-8)
dg= dg %>% ungroup() %>% select(ID, GENE, CHR, POS, MAF, BETA)
don= left_join(don, select(dg, ID, GENE), by= 'ID')
names(dg)= c('ID', 'GENE', 'CHR', 'POS_new', 'MAF', 'BETA')

lims= 250000

don$disc= ifelse(don$pvalue> 5*10**-8, 0, NA)

for (i in rownames(dg)) {
don= mutate(don, disc= ifelse(CHR== as.integer(dg[i, 'CHR']) & POS>= as.integer(dg[i, 'POS_new']) - lims & POS<= as.integer(dg[i, 'POS_new']) + lims, 2, disc))
}

for (i in rownames(ge)) {
don= mutate(don, disc= ifelse(CHR== as.integer(ge[i, 'CHR']) & POS>= as.integer(ge[i, 'pos_ge']) - lims & POS<= as.integer(ge[i, 'pos_ge']) + lims, 1, disc))
}

don= don[order(don$disc, decreasing= F, na.last= T), ]
don$disc= factor(don$disc, levels=c(0, 1, 2), labels=c('Not significant', 'Previous discovery', 'New discovery'))

cols <- c('Not significant'= 'grey', 'Previous discovery'= colorBlindBlack8[4], 'New discovery'= colorBlindBlack8[2])

don$GENE= ifelse(!is.na(don$GENE), don$nearestGene, don$GENE)
ggplot(don) +
    geom_point(data= don, aes(x=BPcum, y= -log10(pvalue), colour= disc), size=0.3) +   # Show all points
theme_cowplot(font_size= 12) + #theme_minimal_hgrid(12, rel_small = -1) + 
#scale_alpha_manual(values= rep(c(1/10, 1/2), 23)) +
scale_colour_manual(values= cols) +
    scale_x_continuous(label = axisdf$CHR, breaks= axisdf$center, expand=c(0,0) ) + # custom X axis
scale_y_continuous(expand= c(0,0)) +
         xlab('Chromosome') +
    ylab('-log10(pvalue)') +
labs(colour= '') +
geom_hline(yintercept= 0, size= 0.5, colour= 'black') +
geom_hline(yintercept= HC, size= 0.5, linetype= 2, colour= '#878787') +
geom_text_repel(data= don, aes(x= BPcum, y= -log10(pvalue), label= GENE), size= 3, hjust= 1, force= 1, vjust= 1, colour= 'black') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 134 of reports/meta_qc.Rmd 
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
indep= fread(snakemake@input[[6]])

indep$nd2P= sapply(strsplit(indep$SP2, ','), length)
indep= inner_join(indep, dg, by= 'CHR')
indep= filter(indep, BP>= POS_new - 1.5*10**6, BP<= POS_new + 1.5*10**6)
indep_df= group_by(indep, GENE) %>% summarize(total= sum(TOTAL), nsig= sum(NSIG), GWS= n(), sug_ev= sum(nd2P), mP= min(P))
indep_df= indep_df[order(indep_df$mP, decreasing=T), ]
indep_df$GENE= factor(indep_df$GENE, levels= indep_df$GENE)
indep_df= filter(indep_df, !grepl('HLA', GENE))

p1= ggplot(data=indep_df, aes(x= GENE, y= GWS)) + 
geom_col(fill=colorBlindBlack8[2], alpha= 0.6) + 
theme_cowplot() +
ylab('# Independent GW significant') +
xlab('Locus') +
coord_flip()

p2= ggplot(data=indep_df, aes(x= GENE, y= total)) + 
geom_col(fill=colorBlindBlack8[4], alpha= 0.6) + 
theme_cowplot() +
ylab('Total # of \n genetic variants in locus') +
xlab('Locus') +
theme(axis.title.y=element_blank(),
        axis.text.y=element_blank(),
        axis.ticks.y=element_blank()) +
coord_flip() 

p3= ggplot(data=indep_df, aes(x= GENE, y= sug_ev / total)) + 
geom_col(fill=colorBlindBlack8[8], alpha= 0.6) + 
theme_cowplot() +
ylab('Proportion of variants with P<1e-5') +
xlab('Locus') +
ylim(0, 1) +
theme(axis.title.y=element_blank(),
        axis.text.y=element_blank(),
        axis.ticks.y=element_blank()) +
coord_flip() 

plot_grid(p1, p2, p3, align = "h", nrow= 1)
R Markdown ggplot2 From line 207 of reports/meta_qc.Rmd 
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
d= mutate(d, maf_tertiles = ntile(MAF, 3))
m1= round(max(d[d$maf_tertiles== 1, 'MAF']), 3)
m2= round(max(d[d$maf_tertiles== 2, 'MAF']), 3)


d$maf_tertiles= factor(d$maf_tertiles, levels=c("1", "2", "3"), labels=c(paste('MAF<', m1), paste(m1,'< MAF >', m2), paste('MAF>', m2)))

df= arrange(d, pvalue) %>% group_by(maf_tertiles) %>% mutate(exp1= -log10(1:length(pvalue)/length(pvalue)))

ggplot(filter(df, pvalue<0.05), aes(exp1, -log10(pvalue), color= maf_tertiles)) +
  geom_point(size= 0.4) +
scale_color_manual(values= colorBlindBlack8[2:4])+ 
  geom_abline(intercept = 0, slope = 1, alpha = .5) +
labs(colour="") +
theme_cowplot(font_size= 12) +
xlab('Expected (-log10(p-value))') +
ylab('Observed (-log10(p-value))') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 263 of reports/meta_qc.Rmd 
292
293
294
295
296
297
298
don= filter(don, pvalue< 1*10**-4)
d1= filter(don, pvalue<5*10**-8)
d1= d1[order(d1$pvalue, decreasing= F), ]
d1$pvalue= format(d1$pvalue, digits= 3)
don$pvalue= format(don$pvalue, digits= 3)

kable(filter(d1, GENE!= '') %>% select(ID, RSID, GENE, TOTALSAMPLESIZE, EAF, BETA, SE, pvalue, Consequence), col.names= c('CHR:POS:REF:EFF', 'RSID', 'Gene', 'N', 'EAF', 'Beta', 'SE', 'P-value', 'Consequence'), digits= 3) #%>% kable_styling(latex_options = c("striped", "scale_down"))
R Markdown From line 292 of reports/meta_qc.Rmd 
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
if (nrow(dg)>1){

(ggplot(dg, aes(MAF, abs(BETA), size= abs(BETA)),  alpha= 0.7) +
geom_point(colour= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
scale_size_continuous(name= 'Absolute Beta', guide= F) +
geom_text_repel(data= dg, aes(label= GENE), guide= F) +
xlab('Minor allele frequency') +
ylab('Absolute effect size')) %>% print()

} else{ 

print('Only one locus identified, check the table.')
plot_comment=''
}
R Markdown From line 315 of reports/meta_qc.Rmd 
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
z= fread(snakemake@input[[5]], header= T, sep= '\t', select= c('MarkerName', 'Effect', 'P-value', 'HetPVal'))

names(z)= c('ID', 'beta', 'pvalue', 'het_pvalue')
z$ID= gsub(':SNP', '', z$ID)
z$ID= gsub(':INDEL', '', z$ID)

z= inner_join(z, dg, by= 'ID')


if (nrow(z)>1) {

plot_comment= 'No pattern between effect size and heterogeneity. Attention should be paid to the top hit.'

z$Direction= ifelse(z$beta> 0, 'Positive', 'Negative')
ggplot(z, aes(-log10(het_pvalue), -log10(pvalue), size= abs(beta)), alpha= 0.7) +
geom_point(colour= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
scale_size_continuous(name= 'Absolute effect size') +
geom_text_repel(data= z, aes(label= GENE), hjust =1, show.legend = FALSE) +
xlab('-log10(Het pvalue)') +
ylab('-log10(Association pvalue)') +
theme(legend.position="bottom")

} else{
print(paste('Pvalue for heterogeneity: ', z$het_pvalue))
#plot_comment= ''
}
R Markdown ggplot2 From line 338 of reports/meta_qc.Rmd 
376
377
378
379
kable(filter(don, (IMPACT== 'HIGH') | (IMPACT== 'MODERATE')) %>% select(ID, RSID, SYMBOL, TOTALSAMPLESIZE, EAF, BETA, SE, pvalue, Consequence), col.names= c('CHR:POS:REF:EFF', 'RSID', 'Gene', 'N', 'EAF', 'Beta', 'SE', 'P-value', 'Consequence'), digits= 3)

x= readLines(snakemake@input[[2]])
x= x[match('Heritability of phenotype 1', x) + 2]
R Markdown From line 376 of reports/meta_qc.Rmd 
393
Ideally, calculate LDscores from our sample (MOBAGENETICS) or from a bigger cohort (UKBIOBANK).
R Markdown From line 393 of reports/meta_qc.Rmd 
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
d= fread(snakemake@input[[3]])

d$pheno1= gsub('.txt.sumstats.gz', '', sapply(strsplit(as.character(d$p1), '/'), tail, 1))
d$pheno2= gsub('.txt.sumstats.gz', '', sapply(strsplit(as.character(d$p2), '/'), tail, 1))

d$rg= ifelse(d$rg> 1, 1, ifelse(d$rg< ( -1), -1, d$rg))

maxy= with(d, ifelse(max(rg + 1.96*se)> 1, max(rg + 1.96*se), 1))
miny=with(d, ifelse(min(rg - 1.96*se)< -1, min(rg - 1.96*se), -1))

ggplot(d, aes(pheno2, rg, colour= pheno2)) +
  geom_point(alpha= 0.5) +
geom_errorbar(aes(ymin= I(rg - 1.96*se) , ymax= (rg + 1.96 * se)), width=.2, position=position_dodge(.9)) +
theme_cowplot(font_size= 9) +
scale_fill_manual(values= colorBlindBlack8[2:4]) +
scale_colour_manual(guide= F, values= colorBlindBlack8[2:4]) +
xlab('Phenotype') +
ylab(paste0('R coefficient [95% CI] \n', pheno)) +
ylim(c(miny, maxy)) +
theme(legend.position= 'none')

link= 'https://drive.google.com/drive/folders/101ErlqwE4_iFwZFCTM0QZUtUVwOoOE1L?usp=sharing'
R Markdown ggplot2 From line 397 of reports/meta_qc.Rmd 
10
11
12
13
14
15
16
17
18
19
20
library(tint)

knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tint'))
options(htmltools.dir.version = FALSE)
pdf.options(useDingbats = TRUE)
#knitr::opts_chunk$set(dpi=300)
knitr::opts_chunk$set(dev = 'png', warning= FALSE, message= FALSE, dpi= 600)



model= ifelse(grepl('rec', snakemake@input[[1]]), 'recessive', 'dominant')
R Markdown From line 10 of reports/nonadditive_qc.Rmd 
31
32
33
34
35
36
37
38
39
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library(ggrepel)
library("data.table")
options(warn=-1)
R Markdown ggplot2 dplyr data.table tidyr cowplot ggrepel knitr kableExtra From line 31 of reports/nonadditive_qc.Rmd 
44
45
46
47
48
49
50
51
52
53
54
55
56
d= fread(snakemake@input[[1]], h= T)
d$MAF= ifelse(d$EAF>0.5,  1 - d$EAF, d$EAF)

d= filter(d, TOTALSAMPLESIZE> 66106)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

ggplot(d, aes(MAF)) +
  geom_density(fill= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
xlab('MAF')

#d= select(d, -MAF)
R Markdown ggplot2 From line 44 of reports/nonadditive_qc.Rmd 
62
63
64
65
ggplot(d, aes(TOTALSAMPLESIZE)) +
  geom_density(fill= colorBlindBlack8[3]) +
theme_cowplot(font_size= 12) +
xlab('Sample size')
R Markdown ggplot2 From line 62 of reports/nonadditive_qc.Rmd 
95
96
97
98
add_model= fread(snakemake@input[[2]])
add_model$CHR= ifelse(add_model$CHR== 'X', '23', add_model$CHR)
add_model$CHR= as.numeric(add_model$CHR)
add_model$pos= round((add_model$pos1 + add_model$pos2) / 2)
R Markdown From line 95 of reports/nonadditive_qc.Rmd 
110
The same number of loci is obtained when using a larger radius (1.5Mb).
R Markdown From line 110 of reports/nonadditive_qc.Rmd 
114
We note that we used a naive approach to identify independent loci. This should be interpreted cautiously. We mapped top genetic variants to the body (TSS or TES) nearest protein coding gene.
R Markdown From line 114 of reports/nonadditive_qc.Rmd 
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
df= arrange(d, pvalue)

#df= group_by(df, CHR, POS) %>% filter(row_number() == 1) %>% ungroup()
df= df[!duplicated(df[, c('CHR', 'POS')]), ]

dg= df %>% arrange(CHR, POS) %>% filter(pvalue< 5*10**-8) %>% group_by(CHR) %>%
		mutate(d=POS-lag(POS, default=-Inf), clumpid=cumsum(d>250000)) %>%
		group_by(CHR, clumpid) %>%
		filter(rank(pvalue, ties.method = "random")==1) %>%
		mutate(GENE= ifelse(SYMBOL=='', RSID, SYMBOL)) %>%
		ungroup()

dg= filter(dg, pvalue< 5*10**-8)
dg= group_by(dg, CHR, POS) %>% filter(row_number()== 1)
#dg$GENE= ifelse(grepl('rs|-', dg$GENE), dg$nearestGene, ifelse(dg$GENE=='', dg$nearestGene, dg$GENE))
dg$GENE= dg$nearestGene

  don <- df %>%
    group_by(CHR)      %>%
    summarise(chr_len= max(POS)) %>%
    mutate(tot= cumsum(as.numeric(chr_len))-chr_len) %>% # Calculate cumulative position of each chromosome
    select(-chr_len) %>%
    left_join(df, ., by= 'CHR') %>%
    arrange(CHR, POS) %>% # Add a cumulative position of each SNP
    mutate( BPcum=POS+tot) %>%
	 ungroup()

  axisdf = don %>% group_by(CHR) %>% summarize(center=( max(BPcum) + min(BPcum) ) / 2 )
  names(axisdf)= c('CHR', 'center')
HC= -log10(5*10**-8)
dg= dg %>% ungroup() %>% select(ID, GENE, CHR, POS, MAF)
don= left_join(don, select(dg, ID, GENE), by= 'ID')
names(dg)= c('ID', 'GENE', 'CHR', 'POS_new', 'MAF')

lims= 250000

don$disc= ifelse(don$pvalue> 5*10**-8, 0, NA)

don= data.frame(don)
dg= data.frame(dg)
add_model= data.frame(add_model)

for (i in rownames(dg)) {
don= mutate(don, disc= ifelse(CHR== as.integer(dg[i, 'CHR']) & POS>= as.integer(dg[i, 'POS_new']) - lims & POS<= as.integer(dg[i, 'POS_new']) + lims, 2, disc))
}

for (i in rownames(add_model)) {
don= mutate(don, disc= ifelse(CHR== as.integer(add_model[i, 'CHR']) & POS>= as.integer(add_model[i, 'pos']) - lims & POS<= as.integer(add_model[i, 'pos']) + lims, 1, disc))
}

don= don[order(don$disc, decreasing= F, na.last= T), ]
don$disc= factor(don$disc, levels=c(0, 1, 2), labels=c('Not significant', 'Additive model discovery', 'New discovery'))

cols <- c('Not significant'= 'grey', 'Additive model discovery'= colorBlindBlack8[4], 'New discovery'= colorBlindBlack8[2])

don$GENE= ifelse(!is.na(don$GENE), don$nearestGene, don$GENE)
ggplot(don) +
    geom_point(data= don, aes(x=BPcum, y= -log10(pvalue), colour= disc), size=0.3) +   # Show all points
theme_cowplot(font_size= 12) + #theme_minimal_hgrid(12, rel_small = -1) + 
#scale_alpha_manual(values= rep(c(1/10, 1/2), 23)) +
scale_colour_manual(values= cols) +
    scale_x_continuous(label = axisdf$CHR, breaks= axisdf$center, expand=c(0,0) ) + # custom X axis
scale_y_continuous(expand= c(0,0)) +
         xlab('Chromosome') +
    ylab('-log10(pvalue)') +
labs(colour= '') +
geom_hline(yintercept= 0, size= 0.5, colour= 'black') +
geom_hline(yintercept= HC, size= 0.5, linetype= 2, colour= '#878787') +
geom_text_repel(data= don, aes(x= BPcum, y= -log10(pvalue), label= GENE), size= 3, hjust= 1, force= 1, vjust= 1, colour= 'black') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 120 of reports/nonadditive_qc.Rmd 
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
d= mutate(d, maf_tertiles = ntile(MAF, 3))
m1= round(max(d[d$maf_tertiles== 1, 'MAF']), 3)
m2= round(max(d[d$maf_tertiles== 2, 'MAF']), 3)


d$maf_tertiles= factor(d$maf_tertiles, levels=c("1", "2", "3"), labels=c(paste('MAF<', m1), paste(m1,'< MAF >', m2), paste('MAF>', m2)))

df= arrange(d, pvalue) %>% group_by(maf_tertiles) %>% mutate(exp1= -log10(1:length(pvalue)/length(pvalue)))

ggplot(filter(df, pvalue<0.05), aes(exp1, -log10(pvalue), color= maf_tertiles)) +
  geom_point(size= 0.4) +
scale_color_manual(values= colorBlindBlack8[2:4])+ 
  geom_abline(intercept = 0, slope = 1, alpha = .5) +
labs(colour="") +
theme_cowplot(font_size= 12) +
xlab('Expected (-log10(p-value))') +
ylab('Observed (-log10(p-value))') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 201 of reports/nonadditive_qc.Rmd 
230
231
232
233
234
235
236
don= filter(don, pvalue< 1*10**-4)
d1= filter(don, pvalue<5*10**-8)
d1= d1[order(d1$pvalue, decreasing= F), ]
d1$pvalue= format(d1$pvalue, digits= 3)
don$pvalue= format(don$pvalue, digits= 3)

kable(filter(d1, GENE!= '') %>% select(ID, RSID, GENE, TOTALSAMPLESIZE, EAF, pvalue, Consequence), col.names= c('CHR:POS:REF:EFF', 'RSID', 'Gene', 'N', 'EAF', 'P-value', 'Consequence'), digits= 3) #%>% kable_styling(latex_options = c("striped", "scale_down"))
R Markdown From line 230 of reports/nonadditive_qc.Rmd 
249
kable(filter(don, (IMPACT== 'HIGH') | (IMPACT== 'MODERATE')) %>% select(ID, RSID, SYMBOL, TOTALSAMPLESIZE, EAF, pvalue, Consequence), col.names= c('CHR:POS:REF:EFF', 'RSID', 'Gene', 'N', 'EAF', 'P-value', 'Consequence'), digits= 3)
R Markdown From line 249 of reports/nonadditive_qc.Rmd 
10
11
12
13
14
15
16
17
18
library(tint)

knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tint'))
options(htmltools.dir.version = FALSE)
pdf.options(useDingbats = TRUE)
#knitr::opts_chunk$set(dpi=300)
knitr::opts_chunk$set(dev = 'png', warning= FALSE, message= FALSE, dpi= 600)
cohort= ifelse(grepl('MOBA', snakemake@input[[1]]), 'MoBa', '23andMe')
pheno= ifelse(grepl('GAraw', snakemake@input[[1]]), 'GA days', 'GA normalized')
R Markdown From line 10 of reports/other_meta.Rmd 
30
31
32
33
34
35
36
37
38
library("ggplot2")
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library("kableExtra")
library(ggrepel)
library("data.table")
options(warn=-1)
R Markdown ggplot2 dplyr data.table tidyr cowplot ggrepel knitr kableExtra From line 30 of reports/other_meta.Rmd 
45
46
47
48
49
50
51
52
53
54
55
d= fread(snakemake@input[[1]], h= T)
d$MAF= ifelse(d$EAF>0.5,  1 - d$EAF, d$EAF)

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

ggplot(d, aes(MAF)) +
  geom_density(fill= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
xlab('MAF')

#d= select(d, -MAF)
R Markdown ggplot2 From line 45 of reports/other_meta.Rmd 
61
62
63
64
ggplot(d, aes(TOTALSAMPLESIZE)) +
  geom_density(fill= colorBlindBlack8[3]) +
theme_cowplot(font_size= 12) +
xlab('Sample size')
R Markdown ggplot2 From line 61 of reports/other_meta.Rmd 
72
73
74
75
ggplot(d, aes(BETA)) +
  geom_density(fill= colorBlindBlack8[2]) +
theme_cowplot(font_size= 12) +
xlab('Beta')
R Markdown ggplot2 From line 72 of reports/other_meta.Rmd 
89
90
**Effective sample size** for binary phenotypes was calculated as:  
$$\frac{2}{(\frac{1}{Ncases} + \frac{1}{Ncontrols})}$$
R Markdown From line 89 of reports/other_meta.Rmd 
105
106
107
108
109
110
111
112
113
114
115
116
ge= data.frame(CHR= c(5, 3, 1, 23, 1), pos_ge= c(157895049, 127881613, 22470407, 115164770, 22470407))


if (grepl('GAraw', snakemake@input[[1]])){
topids= c('1:22462111:A:G', '3:128038373:A:C', '5:157896786:C:T', '23:115184372:A:C', '1:228216997:A:C', '3:123112292:C:T', '3:141147414:C:T', '3:155859113:A:G', '23:131268226:C:T', '2:74207357:A:G', '4:174734471:A:G', '6:32589937:A:G', '6:49559793:G:T', '9:16408826:A:G', '20:62692060:A:C')

} else {
topids= c('1:22414785:G:T', '5:157895049:C:T', '23:115129904:C:T', '1:41955090:A:G', '1:50959262:A:C', '3:14293832:A:G', '3:139004333:A:G', '3:141147414:C:T', '3:155862524:A:G', '3:156697097:A:G', '2:74253326:A:G', '4:55895282:C:T', '4:174739258:A:G', '6:32604898:A:G', '8:75315146:C:G', '9:116935764:C:G')
}
fullmeta= fread(snakemake@input[[3]])
fullmeta= filter(fullmeta, ID %in% topids) %>% select(ID, POS, CHR, BETA, SE, pvalue)
names(fullmeta)= c('ID_f', 'POS_f', 'CHR_f', 'BETA_f', 'SE_f', 'pvalue_f')
R Markdown From line 105 of reports/other_meta.Rmd 
128
The same number of loci is obtained when using a larger radius (1.5Mb).
R Markdown From line 128 of reports/other_meta.Rmd 
132
We note that we used a naive approach to identify independent loci. This should be interpreted cautiously.
R Markdown From line 132 of reports/other_meta.Rmd 
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
df= arrange(d, pvalue)

df= df[!duplicated(df[, c('CHR', 'POS')]), ]

dg= df %>% arrange(CHR, POS) %>% filter(pvalue< 5*10**-8) %>% group_by(CHR) %>%
		mutate(d=POS-lag(POS, default=-Inf), clumpid=cumsum(d>250000)) %>%
		group_by(CHR, clumpid) %>%
		filter(rank(pvalue, ties.method = "random")==1) %>%
		mutate(GENE= nearestGene) %>%
		ungroup()

dg= group_by(dg, CHR, POS) %>% filter(row_number()== 1)


  don <- df %>%
    group_by(CHR)      %>%
    summarise(chr_len= max(POS)) %>%
    mutate(tot= cumsum(as.numeric(chr_len))-chr_len) %>% # Calculate cumulative position of each chromosome
    select(-chr_len) %>%
    left_join(df, ., by= 'CHR') %>%
    arrange(CHR, POS) %>% # Add a cumulative position of each SNP
    mutate( BPcum=POS+tot) %>%
	 ungroup()

  axisdf = don %>% group_by(CHR) %>% summarize(center=( max(BPcum) + min(BPcum) ) / 2 )
  names(axisdf)= c('CHR', 'center')
HC= -log10(5*10**-8)
dg= dg %>% ungroup() %>% select(ID, GENE, CHR, POS, MAF, BETA)
don= left_join(don, select(dg, ID, GENE), by= 'ID')
names(dg)= c('ID', 'GENE', 'CHR', 'POS_new', 'MAF', 'BETA')
don= left_join(don, fullmeta, by= c('CHR'= 'CHR_f'))

lims= 250000

don$disc= ifelse((don$POS> don$POS_f - lims) & (don$POS < don$POS_f + lims), 2, 0)
don= don[order(don$disc, decreasing= T, na.last= T), ]
don= group_by(don, ID) %>% filter(row_number() == 1)
don$disc= ifelse(is.na(don$disc), 0, don$disc)

don= left_join(don, select(dg, CHR, POS_new), by= 'CHR')

don$disc= ifelse(don$disc== 2, 2, ifelse((don$POS> (don$POS_new - lims)) & (don$POS < (don$POS_new + lims)), 1, 0))
don$disc= ifelse(is.na(don$disc), 0, don$disc)
don= don[order(don$disc, decreasing= T, na.last= T), ]
don= group_by(don, ID) %>% filter(row_number() == 1)

don$disc= ifelse(is.na(don$disc), 0, ifelse(don$disc== 1, 2, ifelse(don$disc== 2, 1, 0)))
don= don[order(don$disc, decreasing= F, na.last= T), ]

don$disc= factor(don$disc, levels=c(0, 1, 2), labels=c('Not significant', 'Full meta discovery', 'New discovery'))

cols <- c('Not significant'= 'grey', 'Full meta discovery'= colorBlindBlack8[4], 'New discovery'= colorBlindBlack8[2])



ggplot(don) +
    geom_point(data= don, aes(x=BPcum, y= -log10(pvalue), colour= disc), size=0.3) +   # Show all points
theme_cowplot(font_size= 12) + #theme_minimal_hgrid(12, rel_small = -1) + 
#scale_alpha_manual(values= rep(c(1/10, 1/2), 23)) +
scale_colour_manual(values= cols) +
    scale_x_continuous(label = axisdf$CHR, breaks= axisdf$center, expand=c(0,0) ) + # custom X axis
scale_y_continuous(expand= c(0,0)) +
         xlab('Chromosome') +
    ylab('-log10(pvalue)') +
labs(colour= '') +
geom_hline(yintercept= 0, size= 0.5, colour= 'black') +
geom_hline(yintercept= HC, size= 0.5, linetype= 2, colour= '#878787') +
geom_text_repel(data= don, aes(x= BPcum, y= -log10(pvalue), label= GENE), size= 3, hjust= 1, force= 1, vjust= 1, colour= 'black') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 138 of reports/other_meta.Rmd 
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
d= mutate(d, maf_tertiles = ntile(MAF, 3))
m1= round(max(d[d$maf_tertiles== 1, 'MAF']), 3)
m2= round(max(d[d$maf_tertiles== 2, 'MAF']), 3)


d$maf_tertiles= factor(d$maf_tertiles, levels=c("1", "2", "3"), labels=c(paste('MAF<', m1), paste(m1,'< MAF >', m2), paste('MAF>', m2)))

df= arrange(d, pvalue) %>% group_by(maf_tertiles) %>% mutate(exp1= -log10(1:length(pvalue)/length(pvalue)))

ggplot(filter(df, pvalue<0.05), aes(exp1, -log10(pvalue), color= maf_tertiles)) +
  geom_point(size= 0.4) +
scale_color_manual(values= colorBlindBlack8[2:4])+ 
  geom_abline(intercept = 0, slope = 1, alpha = .5) +
labs(colour="") +
theme_cowplot(font_size= 12) +
xlab('Expected (-log10(p-value))') +
ylab('Observed (-log10(p-value))') +
theme(legend.position= 'bottom') +
guides(colour = guide_legend(override.aes = list(size=3)))
R Markdown ggplot2 From line 214 of reports/other_meta.Rmd 
243
244
245
246
247
248
don= filter(don, pvalue< 1*10**-4)
d1= filter(don, pvalue<5*10**-8)
d1$pvalue= format(d1$pvalue, digits= 3)
don$pvalue= format(don$pvalue, digits= 3)
kable(filter(d1, GENE!= '') %>% select(ID, RSID, GENE, TOTALSAMPLESIZE, EAF, BETA, SE, pvalue), col.names= c('CHR:POS:REF:EFF', 'RSID', 'Gene', 'N', 'EAF', 'Beta', 'SE', 'P-value'), digits= 3)
caption= 'As expected, beta increases with decreasing minor allele frequency.'
R Markdown From line 243 of reports/other_meta.Rmd 
258
259
260
261
262
263
264
265
266
267
268
x= inner_join(fullmeta, d, by= c('ID_f'= 'ID'))


(ggplot(x, aes(BETA_f, BETA, size= abs(BETA)),  alpha= 0.7) +
geom_point(colour= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
scale_size_continuous(name= 'Absolute Beta', guide= F) +
geom_text_repel(data= x, aes(label= RSID), guide= F) +
xlab('Effect size full meta-analysis') +
ylab(paste('Effect size without', cohort))+ 
geom_abline(intercept= 0, slope= 1, linetype= 'dashed', colour= 'grey')) %>% print()
R Markdown From line 258 of reports/other_meta.Rmd 
279
280
281
282
283
284
285
(ggplot(x, aes(-log10(pvalue_f), -log10(pvalue)),  alpha= 0.7) +
geom_point(colour= colorBlindBlack8[4]) +
theme_cowplot(font_size= 12) +
geom_text_repel(data= x, aes(label= RSID), guide= F) +
xlab('-log10(pvalue) full meta-analysis') +
ylab(paste('-log10(pvalue) without', cohort))+ 
geom_abline(intercept= 0, slope= 1, linetype= 'dashed', colour= 'grey')) %>% print()
R Markdown From line 279 of reports/other_meta.Rmd 
11
12
13
14
15
16
17
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['#CHROM', 'POS', 'REF', 'ALT', 'AF_EXCLUDING_1000G'])
	d.columns= ['CHR', 'POS', 'oa', 'ea', 'eaf']
	d= d.loc[((d.eaf> 0.05) & (d.eaf<0.95)), :]
	d['eaf']= np.where(d.oa> d.ea, 1 - d.eaf, d.eaf)
	d= d.sample(n= 1000000)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 11 of reports/Snakefile 
27
28
script:
	'file_level_qc.Rmd'
SnakeMake From line 27 of reports/Snakefile 
38
39
	script:
                'file_level_qc.Rmd'
SnakeMake From line 38 of reports/Snakefile 
49
50
	script:
                'file_level_qc.Rmd'
SnakeMake From line 49 of reports/Snakefile 
60
61
	script:
                'file_level_qc.Rmd'
SnakeMake From line 60 of reports/Snakefile 
70
71
script:
	'all_files_QC.Rmd'
SnakeMake From line 70 of reports/Snakefile 
80
81
script:
        'all_files_QC.Rmd'
SnakeMake From line 80 of reports/Snakefile 
91
92
script:
	'all_files_QC.Rmd'
SnakeMake From line 91 of reports/Snakefile 
102
103
script:
        'all_files_QC.Rmd'
SnakeMake From line 102 of reports/Snakefile 
117
118
script:
	'meta_qc.Rmd'
SnakeMake From line 117 of reports/Snakefile 
129
130
131
132
133
134
135
136
137
138
139
140
141
run:
	df= pd.read_csv(input[0], sep= '\t', header= 0)
	df.sort_values('SNP.PP.H4', ascending= False, inplace= True)
	d= df.groupby('locus').head(1).reset_index()
	df['trait']= input[1].split('pph_')[1].replace('.txt', '')
	d['direction']= np.where((d['z.df1'] > 0) & (d['z.df2'] > 0), 'Positive', np.where((d['z.df1'] < 0) & (d['z.df2'] < 0), 'Negative', 'Opposite'))
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	x['trait']= input[1].split('pph_')[1].replace('.txt', '')
	x= pd.merge(x, d[['snp', 'locus', 'SNP.PP.H4', 'direction']])
	x= x.loc[(x['PP.H0.abf'] != 0) & (x['PP.H1.abf'] != 0) & (x['PP.H2.abf'] != 0) & (x['PP.H0.abf'] != 0) & (x['PP.H4.abf'] != 0), :]
	x.dropna(axis= 0, inplace= True)
	x.to_csv(output[0], sep= '\t', header= True, index= False)
	df.to_csv(output[1], sep= '\t', header= True, index= False)
SnakeMake From line 129 of reports/Snakefile 
151
152
script:
	'coloc.Rmd'
SnakeMake From line 151 of reports/Snakefile 
161
162
163
164
165
166
167
168
169
170
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, compression= 'gzip')
	df= pd.read_csv(input[1], sep= '\t', header= 0)
	df['CHR']= np.where(df.CHR== 'X', '23', df.CHR)
	df['CHR']= df.CHR.astype(str).astype(int)
	d= pd.merge(d, df, on= 'CHR')
	d= d.loc[((d.POS> d.pos1) & (d.POS < d.pos2)), :]
	d.sort_values('pvalue', ascending= True, inplace= True)
	d= d.groupby('nearestGene_y').first()
	d.to_csv(output[0], sep= '\t', header= False, index= False, columns= ['ID'])
SnakeMake From line 161 of reports/Snakefile 
179
180
181
182
183
	shell:
                '''
                grep -f {input[0]} {input[1]} > {output[0]} || true
                touch {output[0]}
                '''
SnakeMake From line 179 of reports/Snakefile 
192
193
194
195
196
shell:
        '''
        grep -f {input[0]} {input[1]} > {output[0]} || true
        touch {output[0]}
        '''
SnakeMake From line 192 of reports/Snakefile 
205
206
207
208
209
shell:
        '''
        grep -f {input[0]} {input[1]} > {output[0]} || true
        touch {output[0]}
        '''
SnakeMake From line 205 of reports/Snakefile 
217
218
219
220
221
        shell:
                '''
		grep -f {input[0]} {input[1]} > {output[0]} || true
		touch {output[0]}
		'''
SnakeMake From line 217 of reports/Snakefile 
230
231
232
233
234
235
236
237
run:
	df_list= list()
	for infile in input:
		d= pd.read_csv(infile, sep= '\t', header= None, names= ['SNP', 'CHR', 'POS', 'EAF', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue', 'STRAND', 'maf'])
		d['cohort']= infile.split('temp/')[1].replace('_topvariants.txt', '')
		df_list.append(d)
	d= pd.concat(df_list)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 230 of reports/Snakefile 
245
246
247
248
249
250
251
252
run:
        df_list= list()
        for infile in input:
                d= pd.read_csv(infile, sep= '\t', header= None, names= ['SNP', 'CHR', 'POS', 'EAF', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue', 'STRAND', 'maf'])
                d['cohort']= infile.split('temp/')[1].replace('_topvariants.txt', '')
                df_list.append(d)
        d= pd.concat(df_list)
        d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 245 of reports/Snakefile 
260
261
262
263
264
265
266
267
run:
        df_list= list()
        for infile in input:
                d= pd.read_csv(infile, sep= '\t', header= None, names= ['SNP', 'CHR', 'POS', 'EAF', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue', 'STRAND', 'maf'])
                d['cohort']= infile.split('temp/')[1].replace('_topvariants.txt', '')
                df_list.append(d)
        d= pd.concat(df_list)
        d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 260 of reports/Snakefile 
275
276
277
278
279
run:
        df_list= list()
        for infile in input:
                d= pd.read_csv(infile, sep= '\t', header= None, names= ['SNP', 'CHR', 'POS', 'EAF', 'N', 'REF', 'EFF', 'BETA', 'SE', 'pvalue', 'STRAND', 'maf'])
                d['cohort']= infile.split('temp/')[1].replace('_topvariants.txt', '')
SnakeMake From line 275 of reports/Snakefile 
292
293
script:
	'forest_plots.Rmd'
SnakeMake From line 292 of reports/Snakefile 
303
304
script:
	'other_meta.Rmd'
SnakeMake From line 303 of reports/Snakefile 
314
315
script:
        'other_meta.Rmd'
SnakeMake From line 314 of reports/Snakefile 
326
327
script:
        'other_meta.Rmd'
SnakeMake From line 326 of reports/Snakefile 
337
338
script:
        'nonadditive_qc.Rmd'
SnakeMake From line 337 of reports/Snakefile 
348
349
script:
        'file_level_qc.Rmd'
SnakeMake From line 348 of reports/Snakefile 
359
360
script:
        'file_level_qc.Rmd'
SnakeMake From line 359 of reports/Snakefile 
370
371
script:
        'file_level_qc.Rmd'
SnakeMake From line 370 of reports/Snakefile 
381
382
script:
        'file_level_qc.Rmd'
SnakeMake From line 381 of reports/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
library(data.table)
library(dplyr)
library(metafor)

funk= function(pheno) {

d_temp= d[d$outcome== pheno, ]

df_list= lapply(c('MT', 'MNT', 'PT'), function(i){


df_temp= d_temp[d_temp$haplotype== i, ]
print(nrow(d_temp))
res.FE= rma(yi= beta, sei= se,  data= df_temp, method= "FE")

df= data.frame(beta= res.FE$beta, se= res.FE$se, pvalue= res.FE$pval, lo95= res.FE$ci.lb, up95= res.FE$ci.ub, het_pvalue= res.FE$QEp, outcome= pheno, haplotype= i)

print(df)

return(df)

})

df= do.call('rbind', df_list)

return(df)

}


moba= fread(snakemake@input[[1]])
decode= fread(snakemake@input[[2]])
hunt= fread(snakemake@input[[3]])

d= rbind(moba, decode)
d= rbind(d, hunt)

df_list= lapply(unique(d$outcome), funk)

x= do.call('rbind', df_list)

df= group_by(d, haplotype, outcome) %>% summarize(n= sum(n))

x= inner_join(x, df, by= c('haplotype', 'outcome'))

fwrite(x, snakemake@output[[1]], sep= '\t')
R dplyr data.table metafor From line 1 of repr_traits_PGS/PGS_fetal_growth_meta.R 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
library(data.table)
library(dplyr)
library(metafor)

funk= function(pheno) {

d_temp= d[d$exposure== pheno, ]

df_list= lapply(c('MT', 'MNT', 'PT'), function(i){


df_temp= d_temp[d_temp$haplotype== i, ]
print(nrow(d_temp))
res.FE= rma(yi= beta, sei= se,  data= df_temp, method= "FE")

df= data.frame(beta= res.FE$beta, se= res.FE$se, pvalue= res.FE$pval, lo95= res.FE$ci.lb, up95= res.FE$ci.ub, het_pvalue= res.FE$QEp, exposure= pheno, haplotype= i)

print(df)

return(df)

})

df= do.call('rbind', df_list)

return(df)

}


moba= fread(snakemake@input[[1]])
decode= fread(snakemake@input[[2]])
hunt= fread(snakemake@input[[3]])

d= rbind(moba, decode)
d= rbind(d, hunt)

df_list= lapply(unique(d$exposure), funk)

x= do.call('rbind', df_list)

df= group_by(d, haplotype, exposure) %>% summarize(n= sum(n))

x= inner_join(x, df, by= c('haplotype', 'exposure'))

fwrite(x, snakemake@output[[1]], sep= '\t')
R dplyr data.table metafor From line 1 of repr_traits_PGS/PGS_repr_pheno_meta.R 
13
14
script:
	'PGS_repr_pheno_meta.R'
SnakeMake From line 13 of repr_traits_PGS/Snakefile 
26
27
script:
        'PGS_fetal_growth_meta.R'
SnakeMake From line 26 of repr_traits_PGS/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
library(data.table)
library(dplyr)
library(DESeq2)
library(tidyverse)

df_list= list()

flist= list.files(snakemake@params[[1]], 'CL', full.names=T)

for (i in 1:length(flist)){
d= fread(flist[i])
cname= unlist(strsplit(flist[i], '/'))[10]
d= select(d, Name, NumReads)

names(d)= c('Name', cname)
df_list[[i]]= d

}

x= df_list %>% reduce(left_join, by = "Name")

cols= data.frame(row.names= colnames(x)[2:7], condition= colnames(x)[2:7], subject= colnames(x)[2:7])

cols$condition= gsub('.txt', '', sapply(strsplit(cols$condition, '-'), tail, 1))
cols$subject= sapply(strsplit(cols$subject, '-'), head, 1)
cts= as.matrix(x[, 2:7])
row.names(cts)= x$Name

dds <- DESeqDataSetFromMatrix(countData = round(cts),
                              colData = cols,
                              design= ~ subject + condition)

dds= DESeq(dds)

res= results(dds, name="condition_unt_vs_dec")

res= data.frame(res)
res$geneid= row.names(res)

fwrite(res, snakemake@output[[1]], sep= '\t')
R tidyverse dplyr data.table From line 1 of stromal_cells/rna_seq_dif.R 
16
17
script:
	'rna_seq_dif.R'
SnakeMake From line 16 of stromal_cells/Snakefile 
28
29
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'ID'])
SnakeMake From line 28 of stromal_cells/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
library("dplyr")
library("knitr")
library("tidyr")
library(cowplot)
library(ggrepel)
library("data.table")
library('showtext')

colorBlindBlack8= c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

x= fread(snakemake@input[[1]])

x$p1= gsub('.txt.sumstats.gz', '', apply(x[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x$p2= gsub('.txt.sumstats.gz', '', apply(x[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))

x1= fread(snakemake@input[[2]])

x1$p1= gsub('.txt.sumstats.gz', '', apply(x1[, 'p1'], 1, function(x) unlist(strsplit(x, 'LDscore/'))[2]))
x1$p2= gsub('.txt.sumstats.gz', '', apply(x1[, 'p2'], 1, function(x) unlist(strsplit(x, 'LDSC/'))[2]))
d= rbind(x, x1)


d$trait= d$p2
d$trait= with(d, ifelse(trait== 'GAraw', 'Maternal gestational duration',
ifelse(trait== 'miscarriage', 'Miscarriage',
                ifelse(trait== 'GA_fetal', 'GA fetal effect',
                ifelse(trait== 'BW_maternal', 'Maternal BW',
                ifelse(trait== 'AFB', 'Age at first birth',
                ifelse(trait== 'AMenarche', 'Age at menarche',
                ifelse(trait== 'AMenopause', 'Age at menopause',
                ifelse(trait== 'NLB', 'Number of live births',
                ifelse(trait== 'Testosterone_fem', 'Testosterone (women)',
                ifelse(trait== 'SHBG_fem', 'SHBG (women)',
                ifelse(trait== 'SHBG_male', 'SHBG (men)',
                ifelse(trait== 'CBAT_fem', 'CBAT (women)',
                ifelse(trait== 'CBAT_male', 'CBAT (men)',
                ifelse(trait== 'Oestradiol_fem', 'Oestradiol (women)',
                ifelse(trait== 'POP', 'Pelvic Organ Prolapse',
                ifelse(trait== 'Testosterone_male', 'Testosterone (men)',
                ifelse(trait== 'leiomyoma_uterus', 'Leiomyoma uterus',
                ifelse(trait== 'BW_fetal', 'Fetal',
                ifelse(trait== 'BW_fetal_effect', 'Fetal only',
                ifelse(trait== 'Preeclampsia', 'Pre-eclampsia',
                ifelse(trait== 'BW_maternal_effect', 'Maternal only',
                ifelse(trait== 'PCOS', 'Polycistic ovary syndrome', 'Endometriosis')))))))))))))))))))))))


d= filter(d, grepl('men', trait), !grepl('women', trait))





fwrite(d, snakemake@output[[1]], sep= '\t')
R dplyr data.table tidyr cowplot ggrepel knitr showtext From line 1 of tables/genetic_correlations_males.R 
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
run:
	d= pd.read_csv(input[0], sep='\t', header= 0)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	x['pheno']= 'Gestational duration'
	ptd= pd.read_csv(input[2], sep= '\t', header= 0)
	ptd['pheno']= 'Preterm delivery'
	postterm= pd.read_csv(input[3], sep= '\t', header= 0)
	postterm['pheno']= 'Post term delivery'
	gID= ['3:156697097:A:G', '5:158058432:G:T']
	d= d.loc[d.ID.isin(gID), :]
	d= pd.concat([x, d])
	d= pd.concat([d, ptd])
	d= pd.concat([d, postterm])
	d.sort_values('ID', inplace= True)
	d.to_csv(output[0], header= True, index= False, sep= '\t')
SnakeMake From line 10 of tables/Snakefile 
34
35
36
37
38
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	d= pd.merge(d[['ID', 'pheno']], x, on= 'ID', how= 'inner')
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 34 of tables/Snakefile 
49
50
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
SnakeMake From line 49 of tables/Snakefile 
67
68
69
70
run:
        d= pd.read_csv(input[1], sep= '\t', header= 0)
        top= pd.read_csv(input[2], sep= '\t', header= 0, usecols= ['ID', 'nearestGene', 'RSID'])
        d= pd.merge(d, top, left_on= 'rsid', right_on= 'RSID')
SnakeMake From line 67 of tables/Snakefile 
82
83
84
85
shell:
	'''
	cp {input[0]} {output[0]}
	'''
SnakeMake From line 82 of tables/Snakefile 
 95
 96
 97
 98
 99
100
101
102
103
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	x['eqtl_data']= 'iPSC'
	d= pd.concat([d, x,])
	df= pd.read_csv(input[2], sep= '\t', header= None, names= ['chr', 'pos1', 'pos2', 'Gene_symbol', 'EID'], usecols= ['Gene_symbol', 'EID'])
	df['EID']= df['EID'].str.split('.').str[0]
	d= pd.merge(d, df, left_on= 'gene', right_on= 'EID')
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 95 of tables/Snakefile 
113
114
script:
	'genetic_correlations_males.R'
SnakeMake From line 113 of tables/Snakefile 
122
123
124
125
126
127
128
129
130
run:
	df_list= list()
	for i in input:
		d= pd.read_csv(i, sep= '\t', header= 0, usecols= ['CHR', 'N'])
		coh= i.split('filtered/')[1].replace('.txt', '')
		df_dict= pd.DataFrame({'cohort': coh, 'N': d.N.max()}, index= [0])
		df_list.append(df_dict)
	d= pd.concat(df_list)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 122 of tables/Snakefile 
138
139
140
141
142
run:
	d= pd.read_csv(input[0], sep= '\t', header= None, names= ['ID', 'beta', 'se', 'pvalue', 'trait'])
	d[['CHR', 'POS', 'REF', 'EFF']]= d.ID.str.split(':', expand= True)
	d['CHR']= np.where(d.CHR== '23', 'X', d.CHR)
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 138 of tables/Snakefile 
151
152
153
154
155
156
157
158
159
160
161
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	d= pd.concat([d, x])
	horm= ['CBAT_fem', 'SHBG_fem', 'Testosterone_fem', 'SHBG_fem_cluster', 'Testosterone_fem_cluster']
	df= d.loc[d.trait.isin(horm), :]
	ivw= df.loc[df.method== 'IVW', :]
	egger= df.loc[df.method== 'MR-Egger', :]
	egger_int= df.loc[np.array(df.index[df.method== 'MR-Egger' ] + 1), :]
	d= pd.concat([ivw, egger, egger_int])
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 151 of tables/Snakefile 
170
171
172
173
174
175
176
177
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['MarkerName', 'HetISq', 'HetPVal'])
	d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	d['REF'], d['EFF']= np.where(d.REF> d.EFF, [d.EFF, d.REF], [d.REF, d.EFF])
	d['ID']= np.where(d.REF> d.EFF, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.EFF + ':' + d.REF, d.CHR.apply(str) + ':' + d.POS.apply(str) + ':' + d.REF + ':' + d.EFF)
	d= pd.merge(d, x, on= 'ID')
	d.to_csv(output[0], sep= '\t', header= True, index= False, columns= ['ID', 'HetISq', 'HetPVal'])
SnakeMake From line 170 of tables/Snakefile 
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	d['Allele1']= d['Allele1'].str.upper()
	d['Allele2']= d['Allele2'].str.upper()
	d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
	d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	d['CHR']= d['CHR'].astype(str).astype(int)
	d['POS']= d['POS'].astype(str).astype(int)
	d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
	d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
	df= d.loc[d.pvalue< 5*10**-8, :]
	df.sort_values(by= 'pvalue', ascending= True, inplace= True)
	df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
	df_list= list()
	for chrom in set(df.CHR):
		d_temp= df.loc[df.CHR== chrom, :]
		positions= d_temp.POS.values
		for pos in positions:
			if pos in d_temp.POS.values:
				df_list.append(d_temp.loc[d_temp.POS== pos, :])
				d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
		else:
			continue
	x= pd.concat(df_list)
	x['pos1']= x.POS - 1.5*10**6
	x['pos2']= x.POS + 1.5*10**6
	x['CHR']= x.CHR.astype(str)
	x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
	x.to_csv(output[0], sep='\t', header= True, index= False, columns= ['CHR', 'pos1', 'pos2'])
SnakeMake From line 7 of top_regions/Snakefile 
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
run:
        d= pd.read_csv(input[0], sep= '\t', compression= 'gzip', usecols= ['CHR', 'POS', 'pvalue', 'nearestGene'])
        df= d.loc[d.pvalue< 5*10**-8, :]
        df.sort_values(by= 'pvalue', ascending= True, inplace= True)
        df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
        df_list= list()
        for chrom in set(df.CHR):
                d_temp= df.loc[df.CHR== chrom, :]
                positions= d_temp.POS.values
                for pos in positions:
                        if pos in d_temp.POS.values:
                                df_list.append(d_temp.loc[d_temp.POS== pos, :])
                                d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
                        else:
                                continue
        x= pd.concat(df_list)
        x['pos1']= x.POS - 1.5*10**6
        x['pos2']= x.POS + 1.5*10**6
        x['CHR']= x.CHR.astype(str)
        x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
        x.to_csv(output[0], sep='\t', header= True, index= False, columns= ['CHR', 'pos1', 'pos2', 'nearestGene'])
SnakeMake From line 43 of top_regions/Snakefile 
72
73
74
75
76
77
78
79
80
81
82
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'EAF', 'TOTALSAMPLESIZE', 'REF', 'EFF', 'RSID', 'ID', 'BETA', 'SE', 'pvalue'])
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	x['CHR']= np.where(x.CHR== 'X', '23', x.CHR)
	x['CHR']= x.CHR.apply(int)
	d= pd.merge(d, x, on= 'CHR')
	d= d.loc[((d.POS>= d.pos1) & (d.POS <= d.pos2)), ]
	d.sort_values('pvalue', ascending= True, inplace= True)
	d= d.groupby('nearestGene').head(1)
	d= d[['CHR', 'POS', 'EAF', 'TOTALSAMPLESIZE', 'REF', 'EFF', 'RSID', 'nearestGene', 'ID', 'BETA', 'SE', 'pvalue']]
	d.to_csv(output[0], sep= '\t', header= True, index= False)
SnakeMake From line 72 of top_regions/Snakefile 
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
run:
	for i in range(2):
		d= pd.read_csv(input[i], sep= '\t', compression= 'gzip', usecols= ['CHR', 'POS', 'pvalue', 'nearestGene'])
		df= d.loc[d.pvalue< 5*10**-8, :]
		df.sort_values(by= 'pvalue', ascending= True, inplace= True)
		df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
		df_list= list()
		for chrom in set(df.CHR):
			d_temp= df.loc[df.CHR== chrom, :]
			positions= d_temp.POS.values
			for pos in positions:
				if pos in d_temp.POS.values:
					df_list.append(d_temp.loc[d_temp.POS== pos, :])
					d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
				else:
					continue
		x= pd.concat(df_list)
		x['pos1']= x.POS - 1.5*10**6
		x['pos2']= x.POS + 1.5*10**6
		x['CHR']= x.CHR.astype(str)
		x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
		x.to_csv(output[i], sep='\t', header= True, index= False, columns= ['CHR', 'pos1', 'pos2', 'nearestGene'])
SnakeMake From line 93 of top_regions/Snakefile 
126
127
128
129
130
131
132
133
134
135
136
137
run:
	for i in range(2):
		d= pd.read_csv(input[i], sep= '\t', header= 0, usecols= ['CHR', 'POS', 'EAF', 'TOTALSAMPLESIZE', 'REF', 'EFF', 'RSID', 'ID', 'pvalue'])
		x= pd.read_csv(input[i+2], sep= '\t', header= 0)
		x['CHR']= np.where(x.CHR== 'X', '23', x.CHR)
		x['CHR']= x.CHR.apply(int)
		d= pd.merge(d, x, on= 'CHR')
		d= d.loc[((d.POS>= d.pos1) & (d.POS <= d.pos2)), ]
		d.sort_values('pvalue', ascending= True, inplace= True)
		d= d.groupby('nearestGene').head(1)
		d= d[['CHR', 'POS', 'EAF', 'TOTALSAMPLESIZE', 'REF', 'EFF', 'RSID', 'nearestGene', 'ID', 'pvalue']]
		d.to_csv(output[i], sep= '\t', header= True, index= False)
SnakeMake From line 126 of top_regions/Snakefile 
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
run:
        d= pd.read_csv(input[0], sep= '\t',usecols= ['CHR', 'POS', 'pvalue'])
        df= d.loc[d.pvalue< 5*10**-8, :]
        df.sort_values(by= 'pvalue', ascending= True, inplace= True)
        df.drop_duplicates(subset= ['CHR', 'POS'], keep= 'first', inplace= True)
        df_list= list()
        for chrom in set(df.CHR):
                d_temp= df.loc[df.CHR== chrom, :]
                positions= d_temp.POS.values
                for pos in positions:
                        if pos in d_temp.POS.values:
                                df_list.append(d_temp.loc[d_temp.POS== pos, :])
                                d_temp= d_temp.loc[(d_temp.POS < pos - (1.5*10**6)) | (d_temp.POS> pos + (1.5 * 10**6)), :]
                        else:
                                continue
        x= pd.concat(df_list)
        x['pos1']= x.POS - 1.5*10**6
        x['pos2']= x.POS + 1.5*10**6
        x['CHR']= x.CHR.astype(str)
        x['CHR']= np.where(x.CHR== '23', 'X', x.CHR)
        x.to_csv(output[0], sep='\t', header= True, index= False, columns= ['CHR', 'POS', 'pos1', 'pos2'])
SnakeMake From line 145 of top_regions/Snakefile 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
import pandas as pd
import numpy as np

d= pd.read_csv(snakemake.input[0], sep= '\t', header=0, compression= 'gzip')
d= d.loc[~d['#chrom'].str.contains('_'), :]
d['a1']= d.alts.str.split(',').str[0]
d['a2']= d.alts.str.split(',').str[1]
d['#chrom']= d['#chrom'].str.replace('chr', '')
d['POS']= np.where(d.ref.str.len() < d.alts.str.len(), d.chromStart, d.chromEnd)
d['ref']= np.where(d.ref.str.len()< d.alts.str.len(), 'I', d.ref)
d['ref']= np.where(d.ref.str.len() > d.alts.str.len(), 'D', d.ref)
d['a1']= np.where(d.ref== 'I', 'D', d.a1)
d['a1']= np.where(d.ref== 'D', 'I', d.a1)
df= d.copy()
df= df.loc[df.a2!= '', :]
d.loc[d.ref > d.a1, ['ref', 'a1']] = d.loc[d.ref > d.a1, ['a1', 'ref']].values

d['ID']= d['#chrom'] + ':' + d['POS'].astype(int).astype(str) + ':' + d.ref + ':' + d.a1
df.loc[df.ref > df.a2, ['ref', 'a2']] = df.loc[df.ref > df.a2, ['a2', 'ref']].values
df['ID']= df['#chrom'] + ':' + df['POS'].astype(int).astype(str) + ':' + df.ref + ':' + df.a2
df= df[['ID', 'name']]
d= d[['ID', 'name']]
d= pd.concat([d, df])

# Read RSIDs from HRC
x= pd.read_csv(snakemake.input[1], sep= '\t', header=0, usecols= ['#CHROM', 'POS', 'ID', 'REF', 'ALT'])
x.columns= ['CHROM', 'POS', 'name', 'REF', 'ALT']
x= x.loc[x.name!= '.', :]

x['CHROM']= np.where(x.CHROM== 'X', '23', x.CHROM)
x['CHROM']= x.CHROM.apply(str)

x.loc[x.REF > x.ALT, ['REF', 'ALT']] = x.loc[x.REF > x.ALT, ['ALT', 'REF']].values
x['ID']= x['CHROM'] + ':' + x['POS'].astype(int).astype(str) + ':' + x.REF + ':' + x.ALT
x= x[['ID', 'name']]
x= x.loc[~x.ID.isin(d.ID), :]

d= pd.concat([d, x])

d.to_csv(snakemake.output[0], sep= '\t', header= True, index= False)
Python Pandas numpy From line 1 of VEP/format_dbSNP.py 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
import pandas as pd
import numpy as np
import re

#d= pd.read_csv(snakemake.input[0], sep= '\t', header= 0)

#d['Allele1']= d['Allele1'].str.upper()
#d['Allele2']= d['Allele2'].str.upper()
#d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
#d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
#d['CHR']= d['CHR'].astype(str).astype(int)
#d['POS']= d['POS'].astype(str).astype(int)
#d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
#d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
#d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
#d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)

#d['CHR']= d['CHR'].astype(str).astype(int)
#d['POS']= d['POS'].astype(str).astype(int)

#d['pvalue']= d['pvalue'].astype(str).astype(float)

#d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
#d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF

#d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]

col_list= ['IMPACT', 'DISTANCE', 'SYMBOL', 'SYMBOL_SOURCE', 'BIOTYPE']
df_list= list()

for vep in pd.read_csv(snakemake.input[1], sep= '\t', header= None, names= ['Variation', 'Location', 'Allele', 'Gene', 'Feature', 'Feature_type', 'Consequence', 'cDNA_position', 'CDS_position', 'Protein_position', 'Amino_acids', 'Codons', 'Existing_variation', 'Extra'], comment= '#', chunksize= 100000):
	for i in col_list:
		vep[i]= vep['Extra'].apply(lambda y: dict([(x.split('=', 1)) for x in re.split(';(?=\w)', y) if x.find('=') > -1])[i] if i in y else '')
	vep= vep[['Variation', 'Location', 'Existing_variation', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'SYMBOL_SOURCE', 'BIOTYPE']]
	vep.columns= ['ID', 'Location', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'SYMBOL_SOURCE', 'BIOTYPE']
	vep['BIOTYPE1']= np.where(vep.BIOTYPE== 'protein_coding', 0, np.where(vep.BIOTYPE.str.contains('pseudo'), 2, 1))
	vep['DISTANCE']= np.where(vep.DISTANCE== '', 0, vep.DISTANCE)
	vep[['chr', 'pos', 'All']]= vep.ID.str.split('_', expand= True)
	vep[['EFF', 'REF']]= vep.All.str.split('/', expand= True)
	vep.loc[vep.REF > vep.EFF, ['REF', 'EFF']] = vep.loc[vep.REF > vep.EFF, ['EFF', 'REF']].values
	vep[['CHR', 'POS']]= vep['Location'].str.split(':', expand= True)
	vep['CHR']= np.where(vep['CHR']== 'X', '23', vep['CHR'])
	vep['ID']= vep.CHR.astype(int).astype(str) + ':' + vep.POS.astype(int).astype(str) + ':' + vep.REF + ':' + vep.EFF
	vep= vep[['ID', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'BIOTYPE', 'BIOTYPE1']]
	vep.sort_values(by= ['BIOTYPE1'], ascending= True, inplace= True)
	vep.drop_duplicates(subset= ['ID'], keep= 'first', inplace= True)
	df_list.append(vep)

vep= pd.concat(df_list)

vep.sort_values(by= ['BIOTYPE1'], ascending= True, inplace= True)
vep.drop_duplicates(subset= ['ID'], keep= 'first', inplace= True)
vep= vep[['ID', 'RSID', 'Gene', 'SYMBOL', 'Consequence', 'IMPACT', 'DISTANCE', 'BIOTYPE']]


d= pd.read_csv(snakemake.input[0], sep= '\t', header= 0)
d['Allele1']= d['Allele1'].str.upper()
d['Allele2']= d['Allele2'].str.upper()
d= d.loc[(d.TOTALSAMPLESIZE> (d['TOTALSAMPLESIZE'].max())/ 2), :]
d[['CHR', 'POS', 'REF','EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
d['CHR']= d['CHR'].astype(str).astype(int)
d['POS']= d['POS'].astype(str).astype(int)
d= d[['CHR', 'POS', 'Allele1', 'Allele2', 'TOTALSAMPLESIZE', 'Freq1', 'Effect', 'StdErr', 'P-value']]
d.columns= ['CHR', 'POS', 'EFF', 'REF', 'TOTALSAMPLESIZE', 'EAF', 'BETA', 'SE', 'pvalue']
d['BETA']=np.where(d.REF > d.EFF, -1* d.BETA, d.BETA)
d['EAF']= np.where(d.REF > d.EFF, 1 - d.EAF, d.EAF)
d['CHR']= d['CHR'].astype(str).astype(int)
d['POS']= d['POS'].astype(str).astype(int)
d['pvalue']= d['pvalue'].astype(str).astype(float)
d.loc[d.REF > d.EFF, ['REF', 'EFF']] = d.loc[d.REF > d.EFF, ['EFF', 'REF']].values
d['ID']= d.CHR.astype(int).astype(str) + ':' + d.POS.astype(int).astype(str) + ':' + d.REF + ':' + d.EFF
d= d.loc[((d.pvalue>0) & (d.pvalue <1)), :]
d= pd.merge(d, vep, on= ['ID'], how= 'left')
d.to_csv(snakemake.output[0], header=True, index= False, sep= '\t')
Python Pandas numpy Variant Effect Predictor (VEP) From line 1 of VEP/format_VEP.py 
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0)
	x= pd.read_csv(input[1], sep= '\t', header= 0)
	x['CHR']= np.where(x['CHR']== '23', 'X', x['CHR'])
	d[['CHR', 'POS', 'REF', 'EFF', 'SNP']]= d['MarkerName'].str.split(':', expand= True)
	d= d.loc[d.SNP== 'SNP', :]
	d['POS2']= d['POS']
	d['CHR']= np.where(d['CHR']== '23', 'X', d['CHR'])
	d['POS']= d['POS'].astype(str).astype(int)
	df_list= list()
	for index, row in x.iterrows():
		temp_df= d.loc[d.CHR== row['CHR'], :]
		temp_df= temp_df.loc[((temp_df.POS >= int(row['pos1'])) & (temp_df.POS <= int(row['pos2']))), :]
		df_list.append(temp_df)
	d= pd.concat(df_list)
	d['Allele']= d['Allele1'].str.upper() + '/' + d['Allele2'].str.upper()
	d['STRAND']= '+'
	d.sort_values(by= ['CHR', 'POS'], inplace= True)
	d.to_csv(output[0], sep= '\t', header= False, index= False, columns= ['CHR', 'POS', 'POS2', 'Allele', 'STRAND'])
SnakeMake From line 11 of VEP/Snakefile 
37
38
shell:
	'/home/pol/software/ensembl-vep/vep -i {input[0]} --check_existing --symbol --biotype --cache -O {output[0]} --offline --force_overwrite'
SnakeMake Variant Effect Predictor (VEP) From line 37 of VEP/Snakefile 
48
49
script:
	'format_VEP.py'
SnakeMake From line 48 of VEP/Snakefile 
58
59
60
61
run:
        d= pd.read_csv(input[0], sep= '\t', header= 0)
        x= pd.read_csv(input[1], sep= '\t', header= 0)
        d= d.loc[~d.geneSymbol.isin(x.name2), :]
SnakeMake From line 58 of VEP/Snakefile 
88
89
90
91
92
93
94
95
96
97
98
99
run:
	d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['MarkerName', 'Allele1'])
	d['CHR']= d.MarkerName.str.split(':').str[0]
	d['end']= d.MarkerName.str.split(':').str[1]
	d['CHR']= d.CHR.astype('str').astype('int')
	d['end']= d.end.astype('str').astype('int')
	d['start']= d.end - 1 
	d['MarkerName']= d.MarkerName.str.replace(':SNP', '')
	d['MarkerName']= d.MarkerName.str.replace(':INDEL', '')
	d.sort_values(by= ['CHR', 'start'], inplace= True)
	d= d[['CHR', 'start', 'end', 'MarkerName']]
	d.to_csv(output[0], sep= '\t', header= False, index= False)
SnakeMake From line 88 of VEP/Snakefile 
108
109
shell:
	'bedtools closest -t all -a {input[0]} -b {input[1]} > {output[0]}'
SnakeMake BEDTools From line 108 of VEP/Snakefile 
118
119
script:
	'format_dbSNP.py'
SnakeMake From line 118 of VEP/Snakefile 
129
130
131
132
133
134
135
136
137
138
139
140
run:
	d= pd.read_csv(input[0], sep= '\t', header=0)
	rs= pd.read_csv(input[1], sep= '\t', header=0)
	d= pd.merge(d, rs, on= 'ID', how= 'left')
	d['RSID']= np.where(pd.isnull(d.RSID), d.name, d.RSID)
	d['RSID']= np.where(d.RSID== '', d.name, d.RSID)
	d['RSID']= np.where(d.RSID== '-', d.name, d.RSID)
	d.drop('name', 1, inplace= True)
	ne= pd.read_csv(input[2], sep= '\t', header= None, names= ['CHR', 'X', 'POS', 'ID', 'c1', 'p1', 'p2', 'nearestGene', 'Ensembl_gene'])
	ne= ne[['ID', 'nearestGene']]
	d= pd.merge(d, ne, on= 'ID', how= 'left')
	d.to_csv(output[0], sep= '\t', header= True, index= False, compression= 'gzip')
SnakeMake From line 129 of VEP/Snakefile 
148
149
150
151
152
153
154
run:
        d= pd.read_csv(input[0], sep= '\t', header=0, usecols= ['ID', 'CHR', 'POS'])
        d['end']= d.POS
        d['start']= d.end - 1
        d.sort_values(by= ['CHR', 'start'], inplace= True)
        d= d[['CHR', 'start', 'end', 'ID']]
        d.to_csv(output[0], sep= '\t', header= False, index= False)
SnakeMake From line 148 of VEP/Snakefile 
163
164
shell:
        'bedtools closest -t all -k 2 -a {input[0]} -b {input[1]} > {output[0]}'
SnakeMake BEDTools From line 163 of VEP/Snakefile 
173
174
175
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['ID', 'nearestGene'])
	ne= pd.read_csv(input[1], sep= '\t', header= None, names= ['CHR', 'X', 'POS', 'ID', 'c1', 'p1', 'p2', 'nearestGene2', 'Ensembl_gene'])
SnakeMake From line 173 of VEP/Snakefile 
188
189
190
run:
	d= pd.read_csv(input[0], sep= '\t', header= 0, usecols= ['MarkerName', 'Allele1', 'Allele2', 'P-value'])
	d= d.loc[d['P-value']< 5e-5, :]
SnakeMake From line 188 of VEP/Snakefile 
207
208
shell:
        '/home/pol/software/ensembl-vep/vep -i {input[0]} --check_existing --symbol --biotype --cache -O {output[0]} --offline --force_overwrite'
SnakeMake Variant Effect Predictor (VEP) From line 207 of VEP/Snakefile 
217
218
script:
        'format_VEP.py'
SnakeMake From line 217 of VEP/Snakefile
ShowHide 387 more snippets with no or duplicated tags.

Login to post a comment if you would like to share your experience with this workflow.

Do you know this workflow well? If so, you can request seller status , and start supporting this workflow.

Free

Created: 1yr ago
Updated: 1yr ago
Maitainers: public
URL: https://github.com/PerinatalLab/metaGWAS
Name: metagwas
Version: v1.0.0
Badge:
workflow icon

Insert copied code into your website to add a link to this workflow.

Downloaded: 0
Copyright: Public Domain
License: MIT License
  • Future updates

Related Workflows

psychip_snakemake
Show Details View Workflow
ENCODE pipeline for histone marks developed for the psychENCODE project
public

public

psychip pipeline is an improved version of the ENCODE pipeline for histone marks developed for the psychENCODE project. The o...
raw sequence reads Alignment Sequence alignment report macs2 ucsc-bedclip bedGraphToBigWig BEDTools BWA Picard SAMtools Snakemake
Free
74
ncov_2
Show Details View Workflow
Near-real time tracking of SARS-CoV-2 in Connecticut
public

public

Repository containing scripts to perform near-real time tracking of SARS-CoV-2 in Connecticut using genomic data. This pipeli...
JSON nextclade Augur Biopython FOCUS Pandas Snakemake bs4 epiweeks geopy matplotlib numpy pycountry pycountry-convert uszipcode
Free
28
cellranger-snakemake-gke
Show Details View Workflow
snakemake workflow to run cellranger on a given bucket using gke.
public

public

A Snakemake workflow for running cellranger on a given bucket using Google Kubernetes Engine. The usage of this workflow ...
macs2 ucsc-bedclip bedGraphToBigWig BEDTools BWA Picard SAMtools Snakemake
Free
74
atlas
Show Details View Workflow
ATLAS - Three commands to start analyzing your metagenome data
public

public

Metagenome-atlas is a easy-to-use metagenomic pipeline based on snakemake. It handles all steps from QC, Assembly, Binning, t...
raw sequence reads Genome assembly Annotation track checkm2 gunc prodigal snakemake-wrapper-utils MEGAHIT Atlas BBMap Biopython BioRuby Bwa-mem2 cd-hit CheckM DAS Diamond eggNOG-mapper v2 MetaBAT 2 Minimap2 MMseqs MultiQC Pandas Picard pyfastx SAMtools SemiBin Snakemake SPAdes SqueezeMeta TADpole VAMB CONCOCT ete3 gtdbtk h5py networkx numpy plotly psutil utils metagenomics
Free
175
rna-seq-star-deseq2
Show Details View Workflow
RNA-seq workflow using STAR and DESeq2
public

public

This workflow performs a differential gene expression analysis with STAR and Deseq2. The usage of this workflow is described ...
raw sequence reads Gene expression profiling ID list snakemake-wrapper-utils biomaRt PCAtools BiocParallel BWA Cutadapt DESeq2 GFFutils MultiQC Pandas SAMtools Snakemake STAR cli stringr tidyverse RNA-Seq
Free
41
dna-seq-gatk-variant-calling
Show Details View Workflow
This Snakemake pipeline implements the GATK best-practices workflow
public

public

This Snakemake pipeline implements the GATK best-practices workflow for calling small germline variants. The usage of thi...
VCF raw sequence reads Variant calling genetic variants gatk rust-bio-tools snakemake-wrapper-utils tabix BCFtools BWA FastQC MultiQC Pandas Picard SAMtools Snakemake Trimmomatic Variant Effect Predictor (VEP) common matplotlib numpy seaborn DNA
Free
57