Workflow Graph

https://github.com/cancerit/cgpbox

This is a repackaging of the cgpbox pipeline to allow for parellization via the rabix executor and descriptions of the tools included by commmon workflow language.



Descriptions of Individual Tools

1. CGP Bam Stats

Generates useful stats on the wildtype and mutant tumour bams that will be used in the pipeline.

base command


bam_stats -i $BAM_MT_TMP -o $BAM_MT_TMP.bas





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  
input(id = "bam", label = "bam", description = "either wildtype or tumour bam", type = "File", secondaryFiles = list(".bai"), prefix = "-i")

)

arguments

arguments = (CCBList(
  
CommandLineBinding(position = 2, prefix = "-o", valueFrom = list("{return $job.inputs.bam.name + \".bas\"}")),
  
CommandLineBinding(position = 3, valueFrom = list("\"&& ls -lR \""))
  
))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob ="\"std.out\""),
output(id = "bam_statistics", label = "bam stat output", 
description = "bam stat output", type = "File",
inheritMetadataFrom = "#bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"*.bas\""))
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "bam-stats", 
label = "CGP bam stats",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "bam_stats",
stdout = "std.out",
inputs = inputs,
arguments = arguments,
outputs = outputs)

Push app to sbg platform

project$app_add("bam-stats", tool)

2. CGP Genotype Check

base command


compareBamGenotypes.pl \
-o /datastore/output/$NAME_WT/genotyped \
-nb $BAM_WT_TMP \
-j /datastore/output/$NAME_WT/genotyped/result.json \
-tb $BAM_MT_TMP \





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  
input(id = "normal_bam", label = "normal_bam", description = "normal wildtype bam", type = "File", prefix = "-nb", stageInput = "link", secondaryFiles = list(".bai")),
  
input(id = "tumour_bam", label = "tumour_bam", description = "mutant tumour bam", type = "File", prefix = "-tb", stageInput = "link", secondaryFiles = list(".bai"))
  
)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "genotype_comparison_results", label = "genotype comparison results", 
description = "genotype comarison results", type = "File",
inheritMetadataFrom = "#tumour_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"genotype_check/*.json\""))
  
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "genotype-check", 
label = "compare bam genotypes",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "compareBamGenotypes.pl -o genotype_check -j genotype_check/genotype_check_result.json",
stdout = "std.out",
inputs = inputs,
argument = arguments,
outputs = outputs)

Make cwl file

write(tool$toJSON(pretty = TRUE), "genotype_check.json")

Push app to sbg platform

project$app_add("genotype-check", tool)

3. Verify Bams

base command


verifyBamHomChk.pl
-b [bam file]
-a [tumour only - from ASCAT tool - called copynumber.caveman.csv ] 
-d 25
-o ./
-j result.json





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  
input(id = "bam", label = "bam", description = "either normal wildtype or tumour mutatantbam", type = "File", prefix = "-b", secondaryFiles = list(".bai")),
  
input(id = "ascat_adjusted_copynumber_csv", label = "ascat_adjusted_copynumber_csv", description = "ascat adjusted for caveman copynumber csv", type = "File", valueFrom = list("{if ($job.inputs.bam.metadata.sample_type == \"Tumor\"){return $self.path;}}"), prefix = "-a")

)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"*.{out}\"")),
  
output(id = "verify_bam_result", label = "verify_bam_results", 
description = "verify bam results", type = "File",
inheritMetadataFrom = "#bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"verify_bam/verify_bam_result.json\""))
  
)

Define Tool Object and Push to Platform

Tool object

need to add some logic for -a adjusted copy number because it is only needed if it is the tumour sample.

-a /datastore/output/${NAME_MT}_vs_${NAME_WT}/ascat/${NAME_MT}.copynumber.caveman.csv

arguments = CCBList(CommandLineBinding(position = 2, prefix = "-a", valueFrom = list("\"if $self.metadata.sample_type = \"Tumor\" then \""))),

{if ($job.inputs.bam.metadata.sample_type == "Tumor") {command_line_arg = "ab.copynumber.caveman.csv";} else {command_line_arg = "";} return command_line_arg}

{if ($job.inputs.bam.metadata.sampleType == "Tumor") {command_line_arg = "ab.copynumber.caveman.csv";} else {command_line_arg = "";} return command_line_arg}

tool <- Tool(
id = "verify_bam", 
label = "verify_bam",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "verifyBamHomChk.pl -d 25 -o verify_bam -j verify_bam/verify_bam_result.json",
stdout = "std.out",
inputs = inputs, arguments = arguments, outputs = outputs)

Push app to sbg platform

project$app_add("verify_bams", tool)

4. ASCAT

Somatic copy number analysis using paired end wholegenome sequencing

https://github.com/cancerit/ascatNgs

base command


ascat.pl
-t [tumour bam]
-n [normal wildtype bam]
-r [genome reference]
-c [cpu number]
-rs 'HUMAN' 
-ra GRCh37 
-pr WGS
-pl ILLUMINA
-sg ascat_SnpGcCorrections.tsv
-q 20
-g L
-o ./ascat_output

Prep ASCAT for caveman


&& perl -ne '@F=(split q{,}, $_)[1,2,3,4]; $F[1]-1; print join("\t",@F)."\n";' < copynumber.caveman.csv > ./normal_cn.bed && perl -ne '@F=(split q{,}, $_)[1,2,3,6]; $F[1]-1; print join("\t",@F)."\n";' < copynumber.caveman.csv > ./tumour_cn.bed





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  
input(id = "tumour_bam", label = "tumour_bam", description = "mutant tumour bam", type = "File", secondaryFiles = list(".bai"), prefix = "-t"),
  
input(id = "normal_bam", label = "normal_bam", description = "normal wildtype bam", type = "File", secondaryFiles = list(".bai"), prefix = "-n"),

input(id = "reference", label = "reference", description = "genome reference", type = "File", secondaryFiles = list(".fai"), prefix = "-r"),

input(id = "cgpbox_reference_files", label = "cgpbox_reference_files", description = "extracted reference files", type = "File...", stageInput = "link")

)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "ascat_statistics", label = "ascat_statistics", 
description = "ascat_statistics", type = "File...",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"ascat_output/*.samplestatistics.txt\"")),
  
output(id = "copynumber_caveman", label = "copynumber_caveman", 
description = "copy_number_caveman", type = "File...",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"ascat_output/*copynumber.caveman.csv\""))
)

Add on scripts

prep_for_caveman_sh = fileDef(name = "prep_for_caveman.sh", content = read_file("prep_for_caveman.sh"))

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "ASCAT", 
label = "ASCAT",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
requirements = requirements(prep_for_caveman_sh),

baseCommand = "CPU=`grep -c ^processor /proc/cpuinfo` && ascat.pl -c $CPU -rs Human -ra NCBI37 -pr WGS -pl ILLUMINA -sg ascat_SnpGcCorrections.tsv -q 20 -g L -o ascat_output",

stdout = "std.out",

inputs = inputs,
  
arguments = arguments,
  
outputs = outputs)

Push app to sbg platform

project$app_add("ASCAT", tool)

5. CGP Pindel

Cancer Genome Project Insertion/Deletion detection pipeline based around Pindel

https://github.com/cancerit/cgpPindel

base command


pindel.pl
-t [tumour mutant bam]
-n [normal wildtype bam]
-r [genome reference]
-c $CPU
-sp HUMAN 
-as GRCh37 
-st WGS
-e NC_007605,hs37d5,GL%
-s pindel_simpleRepeats.bed.gz
-f pindel_genomicRules.lst
-g pindel_human.GRCh37.indelCoding.bed.gz
-u pindel_pindel_np.gff3.gz
-sf pindel_softRules.lst
-b brass-ucscHiDepth_0.01_mrg1000_no_exon_coreChrs.bed.gz
-o ./pindel_output





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  

input(id = "tumour_bam", label = "tumour_bam", description = "tumour bam", type = "File", secondaryFiles = list(".bai"), prefix = "-t"),

input(id = "normal_bam", label = "normal_bam", description = "wildtype bam", type = "File", secondaryFiles = list(".bai"), prefix = "-n"),

input(id = "reference", label = "reference", description = "reference", type = "File", secondaryFiles = list(".fai"), prefix = "-r"),
  
input(id = "cgpbox_reference_files", label = "cgpbox_reference_files", description = "extracted reference files", type = "File...", stageInput = "link")

)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "pindel_germline_bed", label = "pindel_germline_bed", 
description = "pindel_germline_bed output", type = "File",
inheritMetadataFrom = "#tumour_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"pindel_output/*.germline.bed\"")),

output(id = "pindel_flagged_vcf", label = "pindel_flagged_vcf", 
description = "pindel_flagged_vcf", type = "File",
inheritMetadataFrom = "#tumour_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"pindel_output/*.flagged.vcf.gz\""), secondaryFiles = list(".tbi"))
  
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "pindel", 
label = "pindel",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "CPU=`grep -c ^processor /proc/cpuinfo` && pindel.pl -c $CPU -sp HUMAN -as GRCh37 -st WGS -e NC_007605,hs37d5,GL% -s pindel_simpleRepeats.bed.gz -f pindel_genomicRules.lst -g pindel_human.GRCh37.indelCoding.bed.gz -u pindel_pindel_np.gff3.gz -sf pindel_softRules.lst -b brass-ucscHiDepth_0.01_mrg1000_no_exon_coreChrs.bed.gz -o pindel_output",
stdout = "std.out",
inputs = inputs,
arguments = arguments,
outputs = outputs)

Push app to sbg platform

project$app_add("pindel", tool)

6. CaVEMan

CaVEMan actually needs the reference index and not the reference (?). But we can attach the reference index as a secondary file to the reference and find the path that way. Need to ask Erik about this.

It uses outputs from Pindel and ASCAT.

base command


caveman.pl
-tb [tumour mutant bam]
-nb [normal wildtype bam]
-r [genome reference index - genome.fa.fai]
-in [pindel .germline.bed]
-tc [tum.cn.bed - from ASCAT]
-nc [norm.cn.bed - from ASCAT]
-t $CPU
-s HUMAN 
-sa GRCh37 
-st WGS
-st genomic
-ig caveman_ucscHiDepth_0.01_merge1000_no_exon.tsv
-b ./ 
-u ./
-o caveman_output





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(

input(id = "tumour_bam", label = "tumour_bam", description = "mutant tumour bam", type = "File", secondaryFiles = list(".bai"), prefix = "-tb"),

input(id = "normal_bam", label = "normal_bam", description = "normal wildtype bam", type = "File", secondaryFiles = list(".bai"), prefix = "-nb"),
  
input(id = "genome_reference", label = "genome_reference", description = "genome_reference", type = "File", secondaryFiles = list(".fai"), valueFrom = list("$self.path+\".fai\""), prefix = "-r"),

input(id = "pindel_germline_bed", label = "pindel_germline_bed", description = "pindel_germline_bed", type = "File", prefix = "-in"),
  
input(id = "ascat_tumour_cn_bed", label = "ascat_tumour_cn_bed", description = "ascat_tumour_cn_bed", type = "File", prefix = "-tc"),
  
input(id = "ascat_normal_cn_bed", label = "ascat_normal_cn_bed", description = "ascat_normal_cn_bed", type = "File", prefix = "-nc"),
  
input(id = "cgpbox_reference_files", label = "cgpbox_reference_files", description = "extracted reference files", type = "File...", stageInput = "link")

)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "caveman_snv_bed", label = "caveman_snv_bed", 
description = "caveman_snv_bed", type = "File...",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"caveman_output/*bed\"")),
  
output(id = "caveman_all_vcf", label = "caveman_all_vcf", 
description = "caveman_all_vcf", type = "File...",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"caveman_output/*vcf.gz\""), secondaryFiles = list(".tbi")),
  
output(id = "caveman_flagged_muts_vcf", label = "caveman_flagged_muts_vcf", 
description = "caveman_flagged_muts_vcf", type = "File",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script ="\"caveman_output/*flagged.muts.vcf.gz\""), secondaryFiles = list(".tbi"))
  
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "CaVEMan", 
label = "CaVEMan",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), mem(1000)),
baseCommand = "CPU=`grep -c ^processor /proc/cpuinfo` && caveman.pl -t $CPU -s HUMAN -sa GRCh37 -st WGS -st genomic -ig caveman_ucscHiDepth_0.01_merge1000_no_exon.tsv -b ./ -u ./ -o caveman_output",
stdout = "std.out",
inputs = inputs,
arguments = arguments,   
outputs = outputs)

Push app to sbg platform

project$app_add("CaVEMan", tool)

notes

needs these 3 inputs from other tools

-in /datastore/output/${NAME_MT}_vs_${NAME_WT}/pindel/${NAME_MT}_vs_${NAME_WT}.germline.bed  \
-tc $TMP/tum.cn.bed \
-nc $TMP/norm.cn.bed \

Prep ASCAT for caveman

echo -e "CaVEMan prep: `date`"

set -x
ASCAT_CN="/datastore/output/${NAME_MT}_vs_${NAME_WT}/ascat/$NAME_MT.copynumber.caveman.csv"
perl -ne '@F=(split q{,}, $_)[1,2,3,4]; $F[1]-1; print join("\t",@F)."\n";' < $ASCAT_CN > $TMP/norm.cn.bed
perl -ne '@F=(split q{,}, $_)[1,2,3,6]; $F[1]-1; print join("\t",@F)."\n";' < $ASCAT_CN > $TMP/tum.cn.bed
set +x

7. Brass

Breakpoints via assembly - Identifies breaks and attempts to assemble rearrangements.

https://github.com/cancerit/BRASS

base command


brass.pl
-t [tumour bam]
-n [normal wildtype bam]
-g [genome reference]
-ss [ascat *.samplestatistics.txt array]
-j 4
-k 4 
-c $CPU
-s HUMAN 
-as GRCh37 
-pr WGS
-pl ILLUMINA
-d brass-ucscHiDepth_0.01_mrg1000_no_exon_coreChrs.bed.gz
-f brass-brass_np.groups.gz
-g_cache vagrent_Homo_sapiens.GRCh37.75.vagrent.cache.gz
-vi brass-viral.1.1.genomic.fa
-mi brass-all_ncbi_bacteria.20150703
-b brass-hs37d5_500bp_windows.gc.bed.gz
-ct brass-Human.GRCh37.CentTelo.tsv
-o ./brass_output





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

Need to add more inputs here.

inputs = list(
  
input(id = "tumour_bam", label = "tumour_bam", description = "tumour bam", type = "File", stageInput = "link", secondaryFiles = list(".bai"), valueFrom = list("{return $self.name}"), prefix = "-t"),

input(id = "normal_bam", label = "normal_bam", description = "wildtype bam", type = "File", stageInput = "link", secondaryFiles = list(".bai"), valueFrom = list("{return $self.name}"), prefix = "-n"),

input(id = "genome_reference", label = "genome_reference", description = "genome_reference", type = "File", secondaryFiles = list(".fai"), prefix = "-g"),
  
input(id = "bam_stats", label = "bam_stats", description = "bam stats", type = "File...", stageInput = "link"),  
  
input(id = "ascat_sample_statistics", label = "ascat_sample_statistics", description = "ascat_sample_statistics", type = "File...", prefix = "-ss"),
  
input(id = "cgpbox_reference_files", label = "cgpbox_reference_files", description = "extracted reference files", type = "File...", stageInput = "link")
  
)

arguments

arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "brass_sv_bedpe", label = "brass_sv_bedpe", 
description = "brass_sv_bedpe", type = "File",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"brass_output/*bedpe.gz\""), secondaryFiles = list(".tbi")),
  
output(id = "brass_vcf", label = "brass_vcf", 
description = "brass_vcf", type = "File...",
inheritMetadataFrom = "#normal_bam", metadata = list(org = "cgp"),
glob = Expression(engine = "#cwl-js-engine",
script = "\"brass_output/*vcf*\""))
  
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "BRASS", 
label = "BRASS",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "CPU=`grep -c ^processor /proc/cpuinfo` && brass.pl -j 4 -k 4 -c $CPU -s HUMAN -as GRCh37 -pr WGS -pl ILLUMINA -d brass-ucscHiDepth_0.01_mrg1000_no_exon_coreChrs.bed.gz -f brass-brass_np.groups.gz -g_cache vagrent_Homo_sapiens.GRCh37.75.vagrent.cache.gz -vi brass-viral.1.1.genomic.fa -mi brass-all_ncbi_bacteria.20150703 -b brass-hs37d5_500bp_windows.gc.bed.gz -ct brass-Human.GRCh37.CentTelo.tsv -o ./brass_output",
stdout = "std.out", inputs = inputs, arguments = arguments, outputs = outputs)

Push app to sbg platform

project$app_add("BRASS", tool)

8. CGP Annotate VCF

CGP Annotate

base command


AnnotateVcf.pl 
-i [VCF input from pindel or caveman]
-c vagrent_Homo_sapiens.GRCh37.75.vagrent.cache.gz
-o annotated.muts.vcf





docker

The docker conatiner is

cgrlab/cgpbox_dev:develop

App ports

Inputs

inputs = list(
  
input(id = "vcf", label = "vcf", description = "vcf prep annotation", type = "File", secondaryFiles = list(".tbi"), prefix = "-i"),

input(id = "cgpbox_reference_files", label = "cgpbox_reference_files", description = "extracted reference files", type = "File...", stageInput = "link")

)

arguments

arguments = arguments = CCBList(CommandLineBinding(position = 99, valueFrom = list("\"&& ls -lR\"")))

Outputs

outputs = list(

output(id = "std_out", label = "std_out", 
description = "standard output", type = "File...",
metadata = list(org = "cgp"),
glob = "\"std.out\""),
  
output(id = "annotated_vcf", label = "annotated_vcf", 
description = "annotated vcf", type = "File",
inheritMetadataFrom = "#vcf", metadata = list(org = "cgp"),
glob = "\"*vcf*\"")
  
)

Define Tool Object and Push to Platform

Tool object

tool <- Tool(
id = "cgp_annotate_vcf", 
label = "CGP Annotate VCFs",
hints = requirements(docker(pull = "cgrlab/cgpbox_dev:develop"), cpu(1), mem(1000)),
baseCommand = "AnnotateVcf.pl -c vagrent_Homo_sapiens.GRCh37.75.vagrent.cache.gz -o annotated.muts.vcf",
stdout = "std.out",
inputs = inputs,
arguments = arguments,
outputs = outputs)

Make cwl file

write(tool$toJSON(pretty = TRUE), "cgp_annotate_vcf.json")

Push app to sbg platform

project$app_add("cgp_annotate_vcf", tool)

notes


# annotate pindel

AnnotateVcf.pl 
-t 
-c $REF_BASE/vagrent/vagrent.cache.gz
-i /datastore/output/${NAME_MT}_vs_${NAME_WT}/pindel/${NAME_MT}_vs_${NAME_WT}.flagged.vcf.gz
-o /datastore/output/${NAME_MT}_vs_${NAME_WT}/pindel/${NAME_MT}_vs_${NAME_WT}.annot.vcf

# annotate caveman

AnnotateVcf.pl 
-t 
-c $REF_BASE/vagrent/vagrent.cache.gz \
-i /datastore/output/${NAME_MT}_vs_${NAME_WT}/caveman/${NAME_MT}_vs_${NAME_WT}.flagged.muts.vcf.gz
-o /datastore/output/${NAME_MT}_vs_${NAME_WT}/c.annot.muts.vcf

CGC Batch Task Launcher