Trimming, alignment, variant calling, and QC for germline data.

public 1yr ago Version: 1.0.0 0 bookmarks

View Workflow

Help improve this workflow!

This workflow has been published but could be further improved with some additional meta data:

Keyword(s) in categories input, output, operation, topic

You can help improve this workflow by suggesting the addition or removal of keywords, suggest changes and report issues, or request to become a maintainer of the Workflow .

This workflow is designed to take either fastqs or gvcfs as input, and emit a joint-called multi-sample VCF. The pipeline analyzes germline samples only, and does not currently support multiplexed data. It is optimized for deployment on AWS parallelcluster, which can be set up as described here , though it should run without issue on any HPC system or local machine.

Authors

Esha Joshi
Cameron Palmer
Bari Jane Ballew (@bballew)

Usage

If you use this workflow in a paper, don't forget to give credits to the authors by citing the URL of this (original) repository and, if available, its DOI (see above).

Step 1: Obtain a copy of this workflow

Clone this repository to your local system, into the place where you want to perform the data analysis.

 git clone git@gitlab.com:data-analysis5/54gene-wgs-germline.git

Step 2: Configure workflow

Configure the workflow according to your needs via editing the files in the config/ folder. Adjust config.yaml to configure the workflow execution, and manifest.txt to specify your sample setup.

A. Config file

This pipeline currently offers two modes. Please specify the run mode in config.yaml .

full : This mode starts with fastq pairs and emits a joint-called, filtered, multi-sample VCF.
joint_genotyping : This mode starts with gVCFs and runs joint-calling and filtering, emitting a multi-sample VCF. The idea here is that samples can be analyzed in batches in the full run mode, and then the batches can be jointly re-genotyped with this mode.
fastq_qc_only : This mode starts with fastq pairs and performs trimming, then runs FastQC on both pre- and post-trimming fastq pairs, and finally emits a MultiQC report. This is designed to be run initially before a full pipeline run, so that if any lanes need to be dropped or trimming parameters adjusted, it can be identified before spending time and compute on alignment, joint calling, etc.

After joint-calling the samples listed in the manifest, the pipeline will create a new multi-sample VCF where samples have been automatically removed based on the following thresholds sourced from the config:

max_het_ratio (defaults to 2.5): excludes samples with het/hom-alt ratios, as calculated by bcftools stats, that are above the threshold
min_avg_depth (defaults to 20.0): excludes samples with average depth of coverage, as calclated by bcftools stats, that are below the threshold
max_contam (defaults to 0.03): only applied in full runs; excludes samples with contamination estimates, as calculated by verifyBamID, that are above the threshold

B. Manifest file

You will need to provide a headerless, whitespace-delimited manifest file to run the pipeline. For full mode and for fastq_qc_only mode:

Columns: readgroup (should be unique) sample_ID path/to/r1.fastq path/to/r2.fastq
readgroup values should be unique, e.g. sampleID_flowcellID
sample_ID should be the same for all fastq pairs from a single sample, and can be different from the fastq filenames

For joint_genotyping mode:

Columns: sample_ID path/to/file.g.vcf.gz
sample_ID values should be unique, and should correspond to the sample IDs in the gvcfs
gvcfs should be zipped and indexed

Step 3: Install the run-time environment

If needed, install miniconda by following the steps here .

Create a conda environment with, minimally, the dependencies defined in environment.yaml .

# create the env
conda env create -f environment.yaml

Step 4: Execute workflow

Activate the conda environment:

conda activate 54gene-wgs-germline

Test your configuration by performing a dry-run via

snakemake --use-conda -n

Execute the workflow locally via

snakemake --use-conda --cores $N

To run the pipeline in a cluster environment, edit wrapper.sh as needed for your system, and then run via

bash run.sh

Alternatively, you may run snakemake pipelines on a cluster via something like this

snakemake --use-conda --cluster qsub --jobs 100

Step 5: Investigate results

Upon pipeline completion, verify that all steps have completed without error by checking the top-level log WGS_<datestamp>.out . The bottom few lines of the file should contain something like nnn of nnn steps (100%) done . Additional job logs (when run on a cluster) are stored in the logs/ directory.

All pipeline results are stored in the results/ directory.

The hard-filtered, joint-called VCF can be found in results/HaplotypeCaller/filtered/HC_variants.hardfiltered.vcf.gz .

For future joint-calling, the gVCFs are located at results/HaplotypeCaller/called/<sample>_all_chroms.g.vcf.gz .

Deduplicated and post-BQSR bams are found at results/bqsr/<sample>.bam .

Samples that fail the following thresholds are automatically removed from the above file, and the output is placed in results/post_qc_exclusions/samples_excluded.HC_variants.hardfiltered.vcf.gz . The record of sample exclusions, along with reasons for exclusion, is found at results/post_qc_exclusions/exclude_list_with_annotation.tsv . Samples are excluded for at least one of the following reasons. Values listed are defaults, but can be changed in the config.yaml .

Average depth of coverage <20x
Contamination > 3%
Het/Hom ratio > 2.5

The following QC metrics are available:

fastqc at results/fastqc/
Trimming stats via fastp at results/paired_trimmed_reads/
Alignment stats via samtools at results/alignment_stats/
Recalibration stats from bqsr at results/bqsr/
Relatedness via somalier at results/qc/relatedness/
Sample contamination via verifyBamID at results/qc/contamination_check/ - for full runs; not included in joint-genotyping only
Inferred sex via bcftools +guess-ploidy at results/qc/sex_check/
Picard metrics at results/HaplotypeCaller/filtered/
bcftools stats at results/qc/bcftools_stats/
multiqc report at results/multiqc/

A final summary report for the run is available at results/run_summary/ . This report contains overview information, such as samples starting vs. samples emitted into the final VCF, a table of key QC stats, and lists of samples flagged for unexpected relatedness, duplication, or sex discordance.

Step 6: Commit changes

Whenever you change something, don't forget to commit the changes back to your github copy of the repository:

git commit -a
git push

Step 7: Obtain updates from upstream

Whenever you want to synchronize your workflow copy with new developments from upstream, do the following.

Once, register the upstream repository in your local copy: git remote add -f upstream git@github.com:snakemake-workflows/54gene-wgs-germline.git or git remote add -f upstream https://github.com/snakemake-workflows/54gene-wgs-germline.git if you do not have setup ssh keys.
Update the upstream version: git fetch upstream .
Create a diff with the current version: git diff HEAD upstream/master workflow > upstream-changes.diff .
Investigate the changes: vim upstream-changes.diff .
Apply the modified diff via: git apply upstream-changes.diff .
Carefully check whether you need to update the config files: git diff HEAD upstream/master config . If so, do it manually, and only where necessary, since you would otherwise likely overwrite your settings and samples.

Step 8: Contribute back

In case you have also changed or added steps, please consider contributing them back to the original repository:

Fork the original repo to a personal or lab account.
Clone the fork to your local system, to a different place than where you ran your analysis.
Copy the modified files from your analysis to the clone of your fork, e.g., cp -r workflow path/to/fork . Make sure to not accidentally copy config file contents or sample sheets. Instead, manually update the example config files if necessary.
Commit and push your changes to your fork.
Create a pull request against the original repository.

Testing

Test cases are in the subfolder tests . They are automatically executed via continuous integration (TBD).

Unit tests for scripts are found in workflow/scripts/ . They can be executed with pytest or testthat for python and R scripts, respectively.

Test input data, configs, and example manifests for both pipeline modes can be found here . Note that there are a few important caveats when testing.

Somalier doesn't seem to be functional on Mac, so make sure you're in a linux environment or comment out that target from the Snakefile (ugh). (Using a container for the OS would solve this problem....)
Make sure you update the manifests to point to wherever you put the input data
You can run tests locally using the script in that same repo, via bash run_local_test.sh .

Code Snippets

shell:
    "mkdir -p {output.t} && "
    "bwa mem "
    "-K 10000000 -M "
    '-R "@RG\\tCN:54gene\\tID:{params.rg}\\tSM:{params.sm}\\tPL:{params.pl}\\tLB:N/A" '
    "-t {threads} "
    "{input.r} {input.r1} {input.r2} | "
    "samtools sort -@ {params.samtools_threads} -T {output.t} -m {params.samtools_mem}M -o {output.bam} - "

SnakeMake SAMtools BWA From line 43 of rules/align.smk

shell:
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" MarkDuplicates '
    "TMP_DIR={params.t} "
    "REMOVE_DUPLICATES=true "
    "INPUT={params.l} "
    "OUTPUT={output.bam} "
    "METRICS_FILE={output.metrics}"

SnakeMake gatk From line 85 of rules/align.smk

shell:
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" BaseRecalibrator '
    "--tmp-dir {params.t} "
    "-R {input.r} "
    "-I {input.bam} "
    "--known-sites {input.snps} "
    "--known-sites {input.indels} "
    "-O {output.table}"

SnakeMake gatk From line 118 of rules/align.smk

shell:
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" ApplyBQSR '
    "--tmp-dir {params.t} "
    "-R {input.r} "
    "-I {input.bam} "
    "--bqsr-recal-file {input.recal} "
    "-O {output.bam}"

SnakeMake gatk From line 153 of rules/align.smk

shell:
    "samtools stats {input.bam} > {output}"

SnakeMake SAMtools From line 175 of rules/align.smk

shell:
    "ln -s {input.r1} {output.r1} && "
    "ln -s {input.r2} {output.r2}"

SnakeMake From line 21 of rules/fastq_qc.smk

shell:
    "fastqc {input.r1} -d {params.t} --quiet -t {threads} --outdir={params.o} && "
    "fastqc {input.r2} -d {params.t} --quiet -t {threads} --outdir={params.o}"

SnakeMake FastQC From line 47 of rules/fastq_qc.smk

shell:
    "fastp -i {input.r1} -I {input.r2} "
    "-w {threads} "
    "-o {output.r1_paired} -O {output.r2_paired} "
    "-h {output.h} -j {output.j} "
    "--adapter_sequence=AGATCGGAAGAGCACACGTCTGAACTCCAGTCA "
    "--adapter_sequence_r2=AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT "
    "-l 36"

SnakeMake fastp From line 77 of rules/fastq_qc.smk

shell:
    "bcftools norm "
    "-f {input.ref} "
    "-m - "
    "--threads {threads} "
    "{input.vcf} "
    "-Oz -o {output.vcf} && "
    "tabix -p vcf {output.vcf}"

SnakeMake BCFtools tabix From line 24 of rules/filter.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" SelectVariants '
    "--tmp-dir {params.t} "
    "-V {input.vcf} "
    "--select-type SNP "
    "-O {output.vcf}"

SnakeMake gatk From line 53 of rules/filter.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" SelectVariants '
    "--tmp-dir {params.t} "
    "-V {input.vcf} "
    "--select-type INDEL "
    "-O {output.vcf}"

SnakeMake gatk From line 81 of rules/filter.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" VariantFiltration '
    "--tmp-dir {params.t} "
    "-V {input.vcf} "
    '-filter "QD < 2.0" --filter-name "QD2" '
    '-filter "QUAL < 30.0" --filter-name "QUAL30" '
    '-filter "SOR > 3.0" --filter-name "SOR3" '
    '-filter "FS > 60.0" --filter-name "FS60" '
    '-filter "MQ < 40.0" --filter-name "MQ40" '
    '-filter "MQRankSum < -12.5" --filter-name "MQRankSum-12.5" '
    '-filter "ReadPosRankSum < -8.0" --filter-name "ReadPosRankSum-8" '
    "-O {output.vcf}"

SnakeMake gatk From line 109 of rules/filter.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" VariantFiltration '
    "--tmp-dir {params.t} "
    "-V {input.vcf} "
    '-filter "QD < 2.0" --filter-name "QD2" '
    '-filter "QUAL < 30.0" --filter-name "QUAL30" '
    '-filter "FS > 200.0" --filter-name "FS200" '
    '-filter "ReadPosRankSum < -20.0" --filter-name "ReadPosRankSum-20" '
    "-O {output.vcf}"

SnakeMake gatk From line 143 of rules/filter.smk

shell:
    "bcftools view --threads {threads} -r {params.region} -Oz -o {output.vcf} {input.vcf} && "
    "tabix -p vcf {output.vcf}"

SnakeMake BCFtools tabix From line 179 of rules/filter.smk

shell:
    "verifyBamID "
    "--best "
    "--vcf {input.vcf} "
    "--bam {input.bam} "
    "--out {params.d}"

SnakeMake From line 205 of rules/filter.smk

shell:
    'grep -v "^#" {params}*selfSM  > {output}'

SnakeMake From line 220 of rules/filter.smk

shell:
    "bcftools concat -a -Ov {input.snps} {input.indels} | "
    "bcftools sort -T {params.t} -Oz -o {output.vcf} && "
    "tabix -p vcf {output.vcf}"

SnakeMake BCFtools tabix From line 245 of rules/filter.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" CollectVariantCallingMetrics '
    "--TMP_DIR {params.t} "
    "-I {input.calls} "
    "--DBSNP {input.dbsnp} "
    "-SD {input.d} "
    "-O {params.d}"

SnakeMake gatk From line 279 of rules/filter.smk

shell:
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" HaplotypeCaller '
    "--tmp-dir {params.t} "
    "-R {input.r} "
    "-I {input.bam} "
    "-ERC GVCF "
    "-L {input.intervalfile} "
    "-O {output.gvcf} "
    "-G StandardAnnotation "
    "-G StandardHCAnnotation"

SnakeMake gatk From line 35 of rules/HC_calling.smk

shell:
    "bgzip -c {input.gvcf} > {output.gvcf} && "
    "tabix -p vcf {input.gvcf}.gz"

SnakeMake tabix From line 59 of rules/HC_calling.smk

shell:
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" GatherVcfs '
    "-I {params.l} "
    "-O {output}"

SnakeMake gatk From line 94 of rules/HC_calling.smk

shell:
    "tabix -p vcf {input}"

SnakeMake tabix From line 110 of rules/HC_calling.smk

shell:
    "ln -s {input.gvcf} {output.gvcf} && "
    "ln -s {input.index} {output.index}"

SnakeMake From line 14 of rules/HC_joint_geno.smk

shell:
    "n=$(bcftools query -l {input.gvcf});"
    'echo "${{n}}\t{input.gvcf}" > {output}'

SnakeMake From line 39 of rules/HC_joint_geno.smk

shell:
    "for i in {params.mapDir}*.map; do cat $i >> {output}; done"

SnakeMake From line 54 of rules/HC_joint_geno.smk

shell:
    'export _JAVA_OPTIONS="" && '
    "rm -r {params.db} && "
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" GenomicsDBImport '
    "--batch-size {params.batch_size} "
    "--disable-bam-index-caching "
    "--sample-name-map {input.sampleMap} "
    "--genomicsdb-workspace-path {params.db} "
    "-L {input.intervalfile} "
    "--tmp-dir {params.t} "
    "--reader-threads {params.reader_threads} "
    "--genomicsdb-shared-posixfs-optimizations"

SnakeMake gatk From line 126 of rules/HC_joint_geno.smk

shell:
    'export _JAVA_OPTIONS="" && '
    'gatk --java-options "-Xmx{resources.xmx}m {params.java_opts}" GenotypeGVCFs '
    "-R {input.r} "
    "-V gendb://{params.db} "
    "-O {output.vcf} "
    "--tmp-dir {params.t} "
    "--max-alternate-alleles {params.m} "
    "--genomicsdb-max-alternate-alleles {params.g} "
    "--max-genotype-count {params.c} "
    "-stand-call-conf 30 "
    "-G StandardAnnotation "
    "-G StandardHCAnnotation"

SnakeMake gatk From line 170 of rules/HC_joint_geno.smk

shell:
    "bcftools concat -a {input.vcfList} -Ou | "
    "bcftools sort -T {params.t} -Oz -o {output.projectVCF} && "
    "tabix -p vcf {output.projectVCF}"

SnakeMake BCFtools tabix From line 204 of rules/HC_joint_geno.smk

shell:
    "bcftools stats "
    "--af-bins 0.01,0.05,0.1,1 "
    "-F {input.r} "
    "-s- {input.vcf} > {output}"

SnakeMake BCFtools From line 16 of rules/post_calling_qc.smk

shell:
    "plot-vcfstats -P -p {params.d} {input}"

SnakeMake vcfstats From line 52 of rules/post_calling_qc.smk

shell:
    "python workflow/scripts/generate_ped.py {input} {params.prefix}"

SnakeMake From line 68 of rules/post_calling_qc.smk

shell:
    "somalier extract "
    "-d {params.d} "
    "--sites $CONDA_PREFIX/share/somalier/sites.hg38.vcf.gz "
    "-f {input.r} {input.vcf} && "
    "somalier relate --ped {input.ped} -o {params.o} {params.d}/*.somalier"

SnakeMake From line 90 of rules/post_calling_qc.smk

shell:
    "touch {output}"

SnakeMake From line 106 of rules/post_calling_qc.smk

shell:
    "bcftools +guess-ploidy -g hg38 {input.vcf} -v > {output.txt} && "
    "guess-ploidy.py {output.txt} {params.p}"

SnakeMake BCFtools From line 135 of rules/post_calling_qc.smk

shell:
    "python workflow/scripts/create_exclude_list.py {input.b} {params.out} --verify {input.v} -r {params.r} -d {params.d} -c {params.c}"

SnakeMake From line 163 of rules/post_calling_qc.smk

shell:
    "python workflow/scripts/create_exclude_list.py {input.b} {params.out} -r {params.r} -d {params.d}"

SnakeMake From line 183 of rules/post_calling_qc.smk

shell:
    "bcftools view -S ^{input.l} --threads {threads} -Ou {input.v} | "
    "bcftools annotate --threads {threads} --set-id '%CHROM:%POS:%REF:%ALT' -Oz -o {output.v} && "
    "tabix -p vcf {output.v}"

SnakeMake BCFtools tabix From line 200 of rules/post_calling_qc.smk

shell:
    "multiqc --force -o {params.outDir} -n {params.outName} --config {input.mqc_config} {params.inDirs} {params.relatedness}"

SnakeMake MultiQC From line 245 of rules/post_calling_qc.smk

shell:
    "multiqc --force -o {params.outDir} --config {input.mqc_config} -n {params.outName} {params.inDirs} {params.relatedness}"

SnakeMake MultiQC From line 279 of rules/post_calling_qc.smk

shell:
    "multiqc --force -o {params.outDir} --config {input.mqc_config} -n {params.outName} {params.inDirs}"

SnakeMake MultiQC From line 304 of rules/post_calling_qc.smk

shell:
    "bcftools query -l {input} > {output}"

SnakeMake BCFtools From line 14 of rules/report.smk

script:
    "../scripts/run_summary.Rmd"

SnakeMake From line 45 of rules/report.smk

script:
    "../scripts/run_summary.Rmd"

SnakeMake From line 75 of rules/report.smk

script:
    "../scripts/run_initial_summary.Rmd"

SnakeMake From line 101 of rules/report.smk

script:
    "../scripts/combine_benchmarks.R"

SnakeMake From line 129 of rules/report.smk

script:
    "../scripts/benchmarking_report.Rmd"

SnakeMake From line 148 of rules/report.smk

shell:
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.dict resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.alt resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.amb resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.ann resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.bwt resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.pac resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.64.sa resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.fai resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.dbsnp138.vcf resources/ --no-sign-request && "
    "aws s3 cp s3://broad-references/hg38/v0/Homo_sapiens_assembly38.dbsnp138.vcf.idx resources/ --no-sign-request "

SnakeMake From line 46 of rules/resources.smk

shell:
    "date +%s > {output}"

SnakeMake From line 5 of rules/track_time.smk

# Load required packages
knitr::opts_chunk$set(echo = TRUE, fig.width=12, fig.height=8)
require(knitr, quietly = TRUE)
require(ggplot2, quietly=TRUE)
require(RColorBrewer, quietly=TRUE)
require(dplyr, quietly = TRUE)
require(gluedown, quietly = TRUE)
require(gridExtra, quietly = TRUE)
require(lubridate, quietly = TRUE)

# set input variable from snakemake
benchmarks.table <- snakemake@input[["benchmarks"]]

# read in tsv
dataset <- read.table(benchmarks.table, header=TRUE, sep="\t")

# some light date/time manipulation for ease in plotting
dataset$times <- hms(dataset$h.m.s)
d.lub <- hour(dataset$times) + minute(dataset$times)/60 + second(dataset$times)/3600
dataset$walltime <- d.lub

# some upfront unit conversions  & derivations for ease in comprehension
# # convert IO units from MB to GB
dataset$io_in <- dataset$io_in / 1000
dataset$io_out <- dataset$io_out / 1000

# # convert CPU time from seconds to hours
dataset$cpu_time_hrs <- dataset$cpu_time / 3600

# # calculate cpu/walltime ratio
dataset$cpuwall_ratio <- (dataset$cpu_time / dataset$s)

# subset the dataset to get data for rules where there is more than 1
# observation/process to pass to geom_boxplot
subset.df <- dataset %>% group_by(rule) %>% filter(n() > 1)

# subset the dataset to get data points for rules meeting time_threshold
# parameter specified at the rule level
time.param <- snakemake@params[["threshold"]]
time.threshold <- time.param * 60
filtered.df <- dataset %>% group_by(rule) %>% filter(s >= time.threshold)

R Markdown Snakemake ggplot2 dplyr RColorBrewer gridExtra knitr lubridate gluedown From line 15 of scripts/benchmarking_report.Rmd

# plotting function for box/jitter plot where single data points are omitted
# from boxplots
create.boxjitter.plot <- function(df, df.subset, xvar, yvar, xlab, ylab) {
  bd.plot <- ggplot(data=df, aes_string(x=xvar, y=yvar)) +
    geom_boxplot(data = df.subset, aes_string(fill=xvar), show.legend = FALSE)  +
    geom_point(position="identity", show.legend = FALSE, alpha=1/2, size=1) +
    labs(x=xlab, y=ylab) +
    theme_minimal() +
    theme(axis.text.x=element_text(angle = 90, vjust=0.5, hjust=1),
          axis.title = element_text(size = 14),
          axis.text = element_text(size = 12))
  bd.plot
}

create.subset.plot <- function(subdf, xvar, yvar, xlab, ylab, title){
  sb.plot <- ggplot(data = subdf, aes_string(x = xvar, y = yvar)) +
    geom_boxplot(aes_string(fill=xvar), show.legend = FALSE) +
    geom_point(position = "identity", show.legend = FALSE, alpha=1/2, size=1) +
    labs(x=xlab, y=ylab, title=title) +
    theme_minimal() +
    theme(axis.text.x=element_text(angle = 90, vjust=0.5, hjust=1),
          axis.title = element_text(size = 14),
          axis.text = element_text(size = 12))
  sb.plot
}

R Markdown ggplot2 From line 59 of scripts/benchmarking_report.Rmd

inlines <- c(
  md_bold(unique(dataset$rule))
)
md_bullet(inlines)

R Markdown From line 92 of scripts/benchmarking_report.Rmd

# plot walltime for all rules benchmarked
exec.time.all <- create.boxjitter.plot(dataset, subset.df, "rule", "walltime", "Rule", "Total walltime in hours for all rules")
exec.time.all

R Markdown From line 107 of scripts/benchmarking_report.Rmd

max.pss.all <- create.boxjitter.plot(dataset, subset.df, "rule", "max_pss",
                                     "Rule", "PSS (MB) for all rules") +
  scale_y_continuous(n.breaks = 12)
max.pss.all

R Markdown From line 121 of scripts/benchmarking_report.Rmd

max.uss.all <- create.boxjitter.plot(dataset, subset.df, "rule", "max_uss",
                                     "Rule", "USS (MB) for all rules") +
  scale_y_continuous(n.breaks = 12)
max.uss.all

R Markdown From line 130 of scripts/benchmarking_report.Rmd

max.rss.all <- create.boxjitter.plot(dataset, subset.df, "rule", "max_rss",
                                     "Rule", "RSS (MB) for all rules") +
  scale_y_continuous(n.breaks = 12)
max.rss.all

R Markdown From line 139 of scripts/benchmarking_report.Rmd

max.vms.all <- create.boxjitter.plot(dataset, subset.df, "rule", "max_vms",
                                     "Rule", "VMS (MB) for all rules") +
  scale_y_continuous(n.breaks = 12)
max.vms.all

R Markdown From line 148 of scripts/benchmarking_report.Rmd

cat(paste0("Time threshold (minutes): ", time.param))

R Markdown From line 159 of scripts/benchmarking_report.Rmd

# plot the memory usage with boxplots for rules with >1 process otherwise
# just plot the datapoint
max.pss.detail <- create.boxjitter.plot(filtered.df, subset.df, "rule", "max_pss", "Rule", "PSS (MB)") +
  scale_y_continuous(n.breaks = 12)
max.pss.detail

R Markdown From line 170 of scripts/benchmarking_report.Rmd

max.uss.detail <- create.boxjitter.plot(filtered.df, subset.df, "rule", "max_uss", "Rule", "USS (MB)") +
  scale_y_continuous(n.breaks = 12)
max.uss.detail

R Markdown From line 180 of scripts/benchmarking_report.Rmd

max.rss.detail <- create.boxjitter.plot(dataset, subset.df, "rule", "max_rss", "Rule", "RSS (MB)") +
  scale_y_continuous(n.breaks = 12)
max.rss.detail

R Markdown From line 188 of scripts/benchmarking_report.Rmd

max.vms.detail <- create.boxjitter.plot(filtered.df, subset.df, "rule", "max_vms", "Rule", "VMS (MB)") +
  scale_y_continuous(n.breaks = 12)
max.vms.detail

R Markdown From line 196 of scripts/benchmarking_report.Rmd

# plot detailed grid view
mem.df1 <- subset(filtered.df, max_pss > 5000)
mem.df2 <- subset(filtered.df, max_pss < 5000)

max.pss.subset1 <- create.subset.plot(mem.df1, "rule", "max_pss", "Rule", "max_pss (MB)", "Rules using memory >5GB")
max.pss.subset2 <- create.subset.plot(mem.df2, "rule", "max_pss", "Rule", "max_pss (MB)", "Rules using memory <5GB")
grid.arrange(max.pss.subset1, max.pss.subset2, nrow=1)

R Markdown From line 203 of scripts/benchmarking_report.Rmd

# plot the io read
io.read <- create.boxjitter.plot(filtered.df, subset.df, "rule", "io_in", "Rule", "Cumulative GB Read")
io.read

R Markdown From line 220 of scripts/benchmarking_report.Rmd

# plot io write
io.write <- create.boxjitter.plot(filtered.df, subset.df, "rule", "io_out", "Rule", "Cumulative GB Written")
io.write

R Markdown From line 228 of scripts/benchmarking_report.Rmd

# plot detailed grid view
ioin.df1 <- subset(filtered.df, io_in > 10)
ioin.df2 <- subset(filtered.df, io_in < 10)

ioout.df1 <- subset(filtered.df, io_out > 10)
ioout.df2 <- subset(filtered.df, io_out < 10)

ioin.subset1 <- create.subset.plot(ioin.df1, "rule", "io_in", "Rule", "Cumulative GB Written", "Rules with >10GB Read")
ioin.subset2 <- create.subset.plot(ioin.df2, "rule", "io_in", "Rule", "Cumulative GB Written", "Rules with <10GB Read")

ioout.subset1 <- create.subset.plot(ioout.df1, "rule", "io_out", "Rule", "Cumulative GB Written", "Rules with >10GB Written")
ioout.subset2 <- create.subset.plot(ioout.df2, "rule", "io_out", "Rule", "Cumulative GB Written", "Rules with <10GB Written")

grid.arrange(ioin.subset1, ioin.subset2, ioout.subset1, ioout.subset2, nrow=2)

R Markdown From line 235 of scripts/benchmarking_report.Rmd

# plot the CPU time in hours
cpu.time <- create.boxjitter.plot(filtered.df, subset.df, "rule", "cpu_time_hrs", "Rule", "CPU time in hours")
cpu.time

R Markdown From line 257 of scripts/benchmarking_report.Rmd

# plot cpu to walltime ratio (i.e. CPU/s)
cpuwall.ratio <- create.boxjitter.plot(filtered.df, subset.df, "rule", "cpuwall_ratio", "Rule", "CPU/second")
cpuwall.ratio

R Markdown From line 274 of scripts/benchmarking_report.Rmd

library(tidyverse)

# explicitly specify column types to avoid errors when coercing rows 
col.types <- cols(
  s = col_double(),
  `h:m:s` = col_time(format = ""),
  max_rss = col_double(),
  max_vms = col_double(),
  max_uss = col_double(),
  max_pss = col_double(),
  io_in = col_double(),
  io_out = col_double(),
  mean_load = col_double(),
  cpu_time = col_double()
)
# function to read tsv input and append rule name and process
read_benchmarks <- function(filename) {
  read_tsv(filename, col_types=col.types) %>%
    mutate( 
      process = gsub("*.tsv","",basename(filename)),
      rule =  basename(dirname(filename))
      )
}

# run function for all specified snakemake inputs and output benchmarks tsv
dataset <- snakemake@input[["tsv"]] %>%
  lapply(read_benchmarks) %>%
  bind_rows()
head(dataset)
write_tsv(dataset, snakemake@output[['benchmarks']])

R Snakemake tidyverse From line 6 of scripts/combine_benchmarks.R

import argparse
from typing import Dict, List, Optional, Set, Tuple

import pandas as pd


def get_args():
    """Handle command line arguments"""
    parser = argparse.ArgumentParser(
        description="Generates a list of samples to exclude based on previously-run QC metrics."
    )
    parser.add_argument("bstats", type=str, help="bcftools stats file name")
    parser.add_argument("outfile", type=str, help="Output file prefix")
    parser.add_argument(
        "-v",
        "--verify",
        type=str,
        help="concatenated verifyBamID *.selfSM output files",
        default=False,
    )
    parser.add_argument(
        "-r", "--ratio", type=float, help="maximum allowed het/hom_alt ratio", default=2.5
    )
    parser.add_argument(
        "-d", "--depth", type=float, help="minimum allowed average depth", default=20.0
    )
    parser.add_argument(
        "-c", "--contam", type=float, help="maximum allowed contamination", default=0.03
    )
    results = parser.parse_args()
    return (
        results.bstats,
        results.outfile,
        results.verify,
        results.ratio,
        results.depth,
        results.contam,
    )


def read_in_bcftools_PSC(bstats: str) -> pd.DataFrame:
    """Read into a dataframe just the PSC portion of bcftools stats output"""
    skip: list[int] = [i for i, line in enumerate(open(bstats)) if not line.startswith("PSC")]
    return pd.read_csv(
        bstats,
        sep="\t",
        skiprows=skip,
        header=None,
        names=[
            "PSC",
            "id",
            "sample",
            "nRefHom",
            "nNonRefHom",
            "nHets",
            "nTransitions",
            "nTransversions",
            "nIndels",
            "average depth",
            "nSingletons",
            "nHapRef",
            "nHapAlt",
            "nMissing",
        ],
    )


def exclude_high_het_hom(df: pd.DataFrame, r: float) -> pd.DataFrame:
    """Exclude any samples with a het/hom ratio over 2.5"""
    df.loc[:, "het_hom_ratio"] = df["nHets"] / df["nNonRefHom"]
    df_het = df.loc[(df["het_hom_ratio"] > r)].copy()
    if df_het.empty:
        return pd.DataFrame({"sample": [], "exclude_reason": []})
    else:
        df_het.loc[:, "exclude_reason"] = "high_het_hom"
        return df_het[["sample", "exclude_reason"]]


def exclude_low_depth(df: pd.DataFrame, d: float) -> pd.DataFrame:
    """Setting a 20x avg depth cutoff

    This is reported by the lab as well, but this should catch samples
    that are borderline, that were accidentally included despite low
    coverage, or that are from outside collaborators."""
    df_depth = df.loc[df["average depth"] < d].copy()
    if df_depth.empty:
        return pd.DataFrame({"sample": [], "exclude_reason": []})
    else:
        df_depth.loc[:, "exclude_reason"] = "low_depth"
        return df_depth[["sample", "exclude_reason"]]


def read_in_verifybamid(verify: str) -> pd.DataFrame:
    """Requires single *.selfSM files to be concatenated without headers"""
    return pd.read_csv(
        verify,
        sep="\t",
        header=None,
        names=[
            "#SEQ_ID",
            "RG",
            "CHIP_ID",
            "#SNPS",
            "#READS",
            "AVG_DP",
            "FREEMIX",
            "FREELK1",
            "FREELK0",
            "FREE_RH",
            "FREE_RA",
            "CHIPMIX",
            "CHIPLK1",
            "CHIPLK0",
            "CHIP_RH",
            "CHIP_RA",
            "DPREF",
            "RDPHET",
            "RDPALT",
        ],
    )


def exclude_contam(df_v: pd.DataFrame, c: float) -> pd.DataFrame:
    """Setting a 3% contamination threshold"""
    df_contam = df_v.loc[df_v["FREEMIX"] > c].copy()
    if df_contam.empty:
        return pd.DataFrame({"sample": [], "exclude_reason": []})
    else:
        df_contam.loc[:, "exclude_reason"] = "contamination"
        t = df_contam["#SEQ_ID"].str.split(":", expand=True)
        if len(t.columns) == 2:
            df_contam["sample"] = t[1]
        else:
            # if you're running a single sample, verifyBamID doesn't use the sample_self:sample format in this field
            df_contam["sample"] = t[0]
        return df_contam[["sample", "exclude_reason"]]


def combine_exclusions(df_list: list) -> pd.DataFrame:
    """"""
    all_df = pd.concat(df_list).groupby("sample")["exclude_reason"].apply(",".join).reset_index()
    return all_df


if __name__ == "__main__":
    bstats, outfile, verify, r, d, c = get_args()
    df = read_in_bcftools_PSC(bstats)
    exclude1 = exclude_high_het_hom(df, r)
    exclude2 = exclude_low_depth(df, d)
    if verify:
        df_v = read_in_verifybamid(verify)
        exclude3 = exclude_contam(df_v, c)
        exclude_df = combine_exclusions([exclude1, exclude2, exclude3])
    else:
        exclude_df = combine_exclusions([exclude1, exclude2])

    exclude_df.to_csv(outfile + "_with_annotation.tsv", sep="\t", header=None, index=None)
    exclude_df[["sample"]].to_csv(outfile + ".tsv", sep="\t", header=None, index=None)

Python Pandas BCFtools From line 3 of scripts/create_exclude_list.py

import argparse as ap
import re
import shutil
import sys

import pandas as pd


def get_args() -> list:
    """Get command line arguments"""
    parser = ap.ArgumentParser(
        description="Creates a plink-formatted ped from a tsv with subject and sex"
    )
    parser.add_argument("infile", type=str, help="linker filename")
    parser.add_argument(
        "outprefix", type=str, help="output filename prefix (.ped extension will be added)"
    )
    results = parser.parse_args()
    return (results.infile, results.outprefix)


def check_if_ped(infile: str) -> bool:
    """Look for .ped file extension

    This script can either be provided a tsv, which it will
    convert to a plink-style ped, or it can be provided a ped
    which it will simply copy to the expected output name/
    location.
    """
    if re.search(".ped$", infile, re.IGNORECASE):
        return True
    else:
        return False


def read_in_linker(infile: str) -> pd.DataFrame:
    """Read tsv linker file into a dataframe"""
    df = pd.read_table(infile, sep="\t")
    return df


def check_column_number(df: pd.DataFrame):
    """Check for two and only two columns in input tsv"""
    if len(df.columns) != 2:
        raise


def check_column_headers(df: pd.DataFrame):
    """Check for required headers in input tsv"""
    if ["Sample", "Sex"] != list(df.columns)[0:2]:
        raise


def add_ped_columns(df: pd.DataFrame) -> pd.DataFrame:
    """Add dummy columns to conform to plink ped format

    Should end up with a dataframe like so:
    Sample Sample 0 0 [F|M] -9
    """
    df.insert(1, "a", df.loc[:, "Sample"])
    df.insert(2, "b", 0)
    df.insert(2, "c", 0)
    df["d"] = "-9"
    return df


def check_sex_values(df: pd.DataFrame) -> pd.DataFrame:
    """Check for allowed values for sex information

    Input tsv may only have (case-insensitive) f, female,
    m, or male to encode sex.  Missing data can be encoded
    by one of the standard NA values that pandas can deal
    with automatically.  The error message emitted when this
    error is raised lists out the allowed values.
    """
    df["Sex"] = df["Sex"].str.lower()
    if not all(df[~df["Sex"].isna()].isin(["f", "female", "m", "male"])["Sex"]):
        raise


def encode_sex(df: pd.DataFrame) -> pd.DataFrame:
    """Convert sex information to ped encoding

    Missing sex information is converted to 0; male and
    female are converted to 1 and 2 respectively.
    """
    df["Sex"] = df["Sex"].str.lower()
    df["Sex"] = df["Sex"].replace(["f", "female", "m", "male"], [2, 2, 1, 1])
    df = df.fillna(0)
    df["Sex"] = df["Sex"].astype(int)
    return df


if __name__ == "__main__":
    infile, outprefix = get_args()
    if check_if_ped(infile):
        shutil.copyfile(infile, outprefix + ".ped")
        sys.exit(0)
    df = read_in_linker(infile)
    try:
        check_column_number(df)
    except Exception:
        sys.exit("Error: {} is required to have exactly two columns.".format(infile))
    try:
        check_column_headers(df)
    except Exception:
        sys.exit(
            "Error: Correct headers not detected in {} (requires 'Sample' and 'Sex', in that order).".format(
                infile
            )
        )
    try:
        check_sex_values(df)
    except Exception:
        sys.exit(
            "Error: Sex reported in {} must be represented by f, female, m, or male (case insensitive).  Missing data can be represented by any of the following strings: '', '#N/A', '#N/A N/A', '#NA', '-1.#IND', '-1.#QNAN', '-NaN', '-nan', '1.#IND', '1.#QNAN', '<NA>', 'N/A', 'NA', 'NULL', 'NaN', 'n/a', 'nan', or 'null'.".format(
                infile
            )
        )

    df = add_ped_columns(df)
    df = encode_sex(df)
    df.to_csv(outprefix + ".ped", sep="\t", header=False, index=False)

Python Pandas pLink From line 3 of scripts/generate_ped.py

## Load required R packages

library(knitr, quietly = TRUE)
library(kableExtra, quietly = TRUE)
library(ggplot2, quietly = TRUE)
library(RColorBrewer, quietly = TRUE)

R Markdown ggplot2 RColorBrewer knitr kableExtra From line 23 of scripts/run_initial_summary.Rmd

## Load tested functions for this report
source(snakemake@input[["r_resources"]])

R Markdown From line 32 of scripts/run_initial_summary.Rmd

## Configure standard theme information for ggplot2

my.theme <- theme_light() + theme(plot.title = element_text(size = 15, hjust = 0.5),
                                  axis.title = element_text(size = 14),
                                  axis.text = element_text(size = 12),
                                  strip.background = element_blank(),
                                  strip.text = element_text(size = 14, colour = "black"))

R Markdown ggplot2 From line 37 of scripts/run_initial_summary.Rmd

input.subjects <- snakemake@params[["input_samples"]]
fastqc.filename <- snakemake@input[["fastqc"]]
tsv.path <- snakemake@params[["out_prefix"]]
start.time.filename <- snakemake@input[["start_time"]]

R Markdown FastQC From line 47 of scripts/run_initial_summary.Rmd

res <- prepare.initial.subject.tracking.table(input.subjects)
res <- add.fastqc.data(res, fastqc.filename)
res <- res[order(res[, "Subject"]), ]
rownames(res) <- NULL

knitr::kable(res, caption = "Summary of Subject Fate") %>% kableExtra::kable_styling("striped", position = "left", full_width = FALSE)

write.output.table(res, tsv.path)

R Markdown From line 58 of scripts/run_initial_summary.Rmd

start.time <- read.table(start.time.filename, header = FALSE)[1,1]
end.time <- as.integer(format(Sys.time(), "%s"))
elapsed.time <- round((end.time - start.time) / 3600, 2)

R Markdown From line 74 of scripts/run_initial_summary.Rmd

98	sessionInfo()

R Markdown From line 98 of scripts/run_initial_summary.Rmd

## Load required R packages

library(knitr, quietly = TRUE)
library(kableExtra, quietly = TRUE)
library(ggplot2, quietly = TRUE)
library(RColorBrewer, quietly = TRUE)

R Markdown ggplot2 RColorBrewer knitr kableExtra From line 23 of scripts/run_summary.Rmd

## Load tested functions for this report
source(snakemake@input[["r_resources"]])

R Markdown From line 32 of scripts/run_summary.Rmd

## Configure standard theme information for ggplot2

my.theme <- theme_light() + theme(plot.title = element_text(size = 15, hjust = 0.5),
                                  axis.title = element_text(size = 14),
                                  axis.text = element_text(size = 12),
                                  strip.background = element_blank(),
                                  strip.text = element_text(size = 14, colour = "black"))

R Markdown ggplot2 From line 37 of scripts/run_summary.Rmd

input.subjects <- snakemake@params[["input_samples"]]
exclude.reasons.filename <- snakemake@input[["exclude_list"]]
output.subjects.filename <- snakemake@input[["output_subject_list"]]
somalier.run <- snakemake@params[["somalier"]]
somalier.relatedness.filename <- snakemake@input[["relatedness"]]
somalier.sex.filename <- snakemake@input[["sex"]]
fastqc.filename <- snakemake@input[["fastqc"]]
bcftools.stats.filename <- snakemake@input[["bcftools_stats"]]
tsv.path <- snakemake@params[["out_prefix"]]
start.time.filename <- snakemake@input[["start_time"]]
run.mode <- snakemake@params[["run_mode"]]

R Markdown FastQC From line 47 of scripts/run_summary.Rmd

somalier.related.min <- 0.5
somalier.duplicate.min <- 0.95

R Markdown From line 61 of scripts/run_summary.Rmd

res <- prepare.subject.tracking.table(input.subjects, output.subjects.filename, exclude.reasons.filename)

if (!run.mode == "jointgeno"){
  res <- add.fastqc.data(res, fastqc.filename)
}

res <- add.coverage(res, bcftools.stats.filename)
res <- res[order(res[, "Subject"]), ]
rownames(res) <- NULL

knitr::kable(res, caption = "Summary of Subject Fate") %>% kableExtra::kable_styling("striped", position = "left", full_width = FALSE)

write.output.table(res, tsv.path)

R Markdown From line 73 of scripts/run_summary.Rmd

res <- report.sex.discordances(output.subjects.filename, somalier.sex.filename)

if (nrow(res) == 0) {
    cat("No sex discordant subjects were identified.")
} else {
    print(knitr::kable(res, caption = "Sex Discordant Subjects") %>% kableExtra::kable_styling("striped", position = "left", full_width = FALSE))
}

R Markdown From line 99 of scripts/run_summary.Rmd

res <- report.related.subject.pairs(output.subjects.filename, somalier.relatedness.filename, somalier.duplicate.min, 1.0)

if (nrow(res) == 0) {
    cat("No duplicate subjects were detected above the specified relatedness cutoff.")
} else {
    print(knitr::kable(res, caption = "Genetic Duplicates") %>% kableExtra::kable_styling("striped", position = "left", full_width = FALSE))
}

R Markdown From line 117 of scripts/run_summary.Rmd

res <- report.related.subject.pairs(output.subjects.filename, somalier.relatedness.filename, somalier.related.min, somalier.duplicate.min)

if (nrow(res) == 0) {
    cat("No non-identical related subjects were detected within the specified range.")
} else {
    print(knitr::kable(res, caption = "Potentially Related Non-identical Subjects") %>% kableExtra::kable_styling("striped", position = "left", full_width = FALSE))
}

R Markdown From line 137 of scripts/run_summary.Rmd

start.time <- read.table(start.time.filename, header = FALSE)[1,1]
end.time <- as.integer(format(Sys.time(), "%s"))
elapsed.time <- round((end.time - start.time) / 3600, 2)