Skip to content
This repository was archived by the owner on Jul 21, 2020. It is now read-only.

Commit 0ab43b7

Browse files
author
bshifaw
authored
Wdlupdate (#8)
* updated wdl to gatk4.0.5.0 release * realignment_index_bundle variable to json
1 parent ebb6dde commit 0ab43b7

File tree

6 files changed

+197
-25
lines changed

6 files changed

+197
-25
lines changed

‎mutect2-normal-normal.inputs.json‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -34,6 +34,6 @@
3434
"Mutect2NormalNormal.scatter_count": "10",
3535

3636
"##_COMMENT4": "Docker",
37-
"Mutect2NormalNormal.gatk_docker": "broadinstitute/gatk:4.0.4.0"
37+
"Mutect2NormalNormal.gatk_docker": "broadinstitute/gatk:4.0.5.0"
3838
}
3939

‎mutect2-normal-normal.wdl‎

Lines changed: 5 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -32,13 +32,12 @@ workflow Mutect2NormalNormal {
3232
Array[File] bams
3333
Array[File] bais
3434
File? pon
35-
File? pon_index
3635
File? gnomad
37-
File? gnomad_index
3836
File? variants_for_contamination
39-
File? variants_for_contamination_index
4037
Boolean? run_orientation_bias_filter
4138
Array[String]? artifact_modes
39+
File? realignment_index_bundle
40+
String? realignment_extra_args
4241
String? m2_extra_args
4342
String? m2_extra_filtering_args
4443
Boolean? make_bamout
@@ -68,15 +67,14 @@ workflow Mutect2NormalNormal {
6867
normal_bam = normal_bam,
6968
normal_bai = normal_bai,
7069
pon = pon,
71-
pon_index = pon_index,
7270
scatter_count = scatter_count,
7371
gnomad = gnomad,
74-
gnomad_index = gnomad_index,
7572
variants_for_contamination = variants_for_contamination,
76-
variants_for_contamination_index = variants_for_contamination_index,
7773
run_orientation_bias_filter = run_orientation_bias_filter,
7874
preemptible_attempts = preemptible_attempts,
7975
artifact_modes = artifact_modes,
76+
realignment_index_bundle = realignment_index_bundle,
77+
realignment_extra_args = realignment_extra_args,
8078
m2_extra_args = m2_extra_args,
8179
m2_extra_filtering_args = m2_extra_filtering_args,
8280
gatk_override = gatk_override,
@@ -155,6 +153,7 @@ task CountFalsePositives {
155153

156154
runtime {
157155
docker: gatk_docker
156+
bootDiskSizeGb: 12
158157
memory: "5 GB"
159158
disks: "local-disk " + 500 + " HDD"
160159
}

‎mutect2.exome.inputs.json‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
{
22
"##_COMMENT1": "Runtime",
33
"##Mutect2.oncotator_docker": "(optional) String?",
4-
"Mutect2.gatk_docker": "broadinstitute/gatk:4.0.4.0",
4+
"Mutect2.gatk_docker": "broadinstitute/gatk:4.0.5.0",
55

66
"##_COMMENT2": "Workflow options",
77
"Mutect2.intervals": "gs://gatk-best-practices/somatic-b37/whole_exome_agilent_1.1_refseq_plus_3_boosters.Homo_sapiens_assembly19.baits.interval_list",
@@ -28,6 +28,7 @@
2828
"Mutect2.gnomad_index": "gs://gatk-best-practices/somatic-b37/af-only-gnomad.raw.sites.vcf.idx",
2929
"Mutect2.variants_for_contamination": "gs://gatk-best-practices/somatic-b37/small_exac_common_3.vcf",
3030
"Mutect2.variants_for_contamination_index": "gs://gatk-best-practices/somatic-b37/small_exac_common_3.vcf.idx",
31+
"##Mutect2.realignment_index_bundle": "File? (optional)",
3132

3233
"##_COMMENT5": "Secondary resources",
3334
"Mutect2.onco_ds_tar_gz": "gs://gatk-best-practices/somatic-b37/oncotator_v1_ds_April052016.tar.gz",

‎mutect2.wdl‎

Lines changed: 101 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -41,6 +41,7 @@
4141
## onco_ds_tar_gz, default_config_file: Oncotator datasources and config file
4242
## sequencing_center, sequence_source: metadata for Oncotator
4343
## filter_oncotator_maf: Whether the MAF generated by oncotator should have the filtered variants removed. Default: true
44+
## realignment_index_bundle: resource for FilterAlignmentArtifacts, which runs if and only if it is specified. Generated by BwaMemIndexImageCreator.
4445
##
4546
## Outputs :
4647
## - One VCF file and its index with primary filtering applied; secondary filtering and functional annotation if requested; a bamout.bam
@@ -73,6 +74,8 @@ workflow Mutect2 {
7374
File? gnomad_index
7475
File? variants_for_contamination
7576
File? variants_for_contamination_index
77+
File? realignment_index_bundle
78+
String? realignment_extra_args
7679
Boolean? run_orientation_bias_filter
7780
Boolean run_ob_filter = select_first([run_orientation_bias_filter, false])
7881
Array[String]? artifact_modes
@@ -115,6 +118,8 @@ workflow Mutect2 {
115118
String oncotator_docker_or_default = select_first([oncotator_docker, "broadinstitute/oncotator:1.9.9.0"])
116119
Boolean? filter_oncotator_maf
117120
Boolean filter_oncotator_maf_or_default = select_first([filter_oncotator_maf, true])
121+
Boolean? filter_funcotations
122+
Boolean filter_funcotations_or_default = select_first([filter_funcotations, true])
118123
String? oncotator_extra_args
119124

120125
Int? preemptible_attempts
@@ -302,8 +307,26 @@ workflow Mutect2 {
302307
}
303308
}
304309
310+
if (defined(realignment_index_bundle)) {
311+
File realignment_filter_input = select_first([FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
312+
File realignment_filter_input_idx = select_first([FilterByOrientationBias.filtered_vcf_index, Filter.filtered_vcf_index])
313+
call FilterAlignmentArtifacts {
314+
input:
315+
gatk_override = gatk_override,
316+
bam = tumor_bam,
317+
bai = tumor_bai,
318+
realignment_index_bundle = select_first([realignment_index_bundle]),
319+
realignment_extra_args = realignment_extra_args,
320+
gatk_docker = gatk_docker,
321+
compress = compress,
322+
output_name = filtered_name,
323+
input_vcf = realignment_filter_input,
324+
input_vcf_idx = realignment_filter_input_idx
325+
}
326+
}
327+
305328
if (run_oncotator_or_default) {
306-
File oncotate_vcf_input = select_first([FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
329+
File oncotate_vcf_input = select_first([FilterAlignmentArtifacts.filtered_vcf, FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
307330
call oncotate_m2 {
308331
input:
309332
m2_vcf = oncotate_vcf_input,
@@ -323,8 +346,8 @@ workflow Mutect2 {
323346
}
324347
325348
if (run_funcotator_or_default) {
326-
File funcotate_vcf_input = select_first([FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
327-
File funcotate_vcf_input_index = select_first([FilterByOrientationBias.filtered_vcf_index, Filter.filtered_vcf_index])
349+
File funcotate_vcf_input = select_first([FilterAlignmentArtifacts.filtered_vcf, FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
350+
File funcotate_vcf_input_index = select_first([FilterAlignmentArtifacts.filtered_vcf_index, FilterByOrientationBias.filtered_vcf_index, Filter.filtered_vcf_index])
328351
call Funcotate {
329352
input:
330353
m2_vcf = funcotate_vcf_input,
@@ -341,15 +364,16 @@ workflow Mutect2 {
341364
annotation_defaults = annotation_defaults,
342365
annotation_overrides = annotation_overrides,
343366
gatk_docker = gatk_docker,
344-
gatk_override = gatk_override
367+
gatk_override = gatk_override,
368+
filter_funcotations = filter_funcotations_or_default
345369
}
346370
}
347371
348372
output {
349373
File unfiltered_vcf = MergeVCFs.merged_vcf
350374
File unfiltered_vcf_index = MergeVCFs.merged_vcf_index
351-
File filtered_vcf = select_first([FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
352-
File filtered_vcf_index = select_first([FilterByOrientationBias.filtered_vcf_index, Filter.filtered_vcf_index])
375+
File filtered_vcf = select_first([FilterAlignmentArtifacts.filtered_vcf, FilterByOrientationBias.filtered_vcf, Filter.filtered_vcf])
376+
File filtered_vcf_index = select_first([FilterAlignmentArtifacts.filtered_vcf_index, FilterByOrientationBias.filtered_vcf_index, Filter.filtered_vcf_index])
353377
File? contamination_table = CalculateContamination.contamination_table
354378

355379
File? oncotated_m2_maf = oncotate_m2.oncotated_m2_maf
@@ -358,6 +382,7 @@ workflow Mutect2 {
358382
File? preadapter_detail_metrics = CollectSequencingArtifactMetrics.pre_adapter_metrics
359383
File? bamout = MergeBamOuts.merged_bam_out
360384
File? bamout_index = MergeBamOuts.merged_bam_out_index
385+
File? maf_segments = CalculateContamination.maf_segments
361386
}
362387
}
363388

@@ -400,6 +425,7 @@ task SplitIntervals {
400425

401426
runtime {
402427
docker: gatk_docker
428+
bootDiskSizeGb: 12
403429
memory: machine_mem + " MB"
404430
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
405431
preemptible: select_first([preemptible_attempts, 10])
@@ -481,6 +507,7 @@ task M2 {
481507

482508
runtime {
483509
docker: gatk_docker
510+
bootDiskSizeGb: 12
484511
memory: machine_mem + " MB"
485512
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
486513
preemptible: select_first([preemptible_attempts, 10])
@@ -529,6 +556,7 @@ task MergeVCFs {
529556

530557
runtime {
531558
docker: gatk_docker
559+
bootDiskSizeGb: 12
532560
memory: machine_mem + " MB"
533561
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
534562
preemptible: select_first([preemptible_attempts, 10])
@@ -576,12 +604,13 @@ task MergeBamOuts {
576604
577605
gatk --java-options "-Xmx${command_mem}m" SortSam -I unsorted.out.bam \
578606
-O ${output_vcf_name}.out.bam \
579-
--SORT_ORDER coordinate
580-
gatk --java-options "-Xmx${command_mem}m" BuildBamIndex -I ${output_vcf_name}.out.bam
607+
--SORT_ORDER coordinate -VALIDATION_STRINGENCY LENIENT
608+
gatk --java-options "-Xmx${command_mem}m" BuildBamIndex -I ${output_vcf_name}.out.bam -VALIDATION_STRINGENCY LENIENT
581609
>>>
582610

583611
runtime {
584612
docker: gatk_docker
613+
bootDiskSizeGb: 12
585614
memory: machine_mem + " MB"
586615
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
587616
preemptible: select_first([preemptible_attempts, 10])
@@ -624,6 +653,7 @@ task CollectSequencingArtifactMetrics {
624653

625654
runtime {
626655
docker: gatk_docker
656+
bootDiskSizeGb: 12
627657
memory: machine_mem + " MB"
628658
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
629659
preemptible: select_first([preemptible_attempts, 10])
@@ -676,6 +706,7 @@ task CalculateContamination {
676706

677707
runtime {
678708
docker: gatk_docker
709+
bootDiskSizeGb: 12
679710
memory: command_mem + " MB"
680711
disks: "local-disk " + select_first([disk_space, 100]) + " HDD"
681712
preemptible: select_first([preemptible_attempts, 10])
@@ -729,6 +760,7 @@ task Filter {
729760

730761
runtime {
731762
docker: gatk_docker
763+
bootDiskSizeGb: 12
732764
memory: machine_mem + " MB"
733765
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
734766
preemptible: select_first([preemptible_attempts, 10])
@@ -779,6 +811,61 @@ task FilterByOrientationBias {
779811

780812
runtime {
781813
docker: gatk_docker
814+
bootDiskSizeGb: 12
815+
memory: command_mem + " MB"
816+
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
817+
preemptible: select_first([preemptible_attempts, 10])
818+
cpu: select_first([cpu, 1])
819+
}
820+
821+
output {
822+
File filtered_vcf = "${output_vcf}"
823+
File filtered_vcf_index = "${output_vcf_index}"
824+
}
825+
}
826+
827+
task FilterAlignmentArtifacts {
828+
#input
829+
File? gatk_override
830+
File input_vcf
831+
File input_vcf_idx
832+
File bam
833+
File bai
834+
String output_name
835+
Boolean compress
836+
String output_vcf = output_name + if compress then ".vcf.gz" else ".vcf"
837+
String output_vcf_index = output_vcf + if compress then ".tbi" else ".idx"
838+
File realignment_index_bundle
839+
String? realignment_extra_args
840+
841+
# runtime
842+
String gatk_docker
843+
Int? mem
844+
Int? preemptible_attempts
845+
Int? disk_space
846+
Int? cpu
847+
Boolean use_ssd = false
848+
849+
# Mem is in units of GB but our command and memory runtime values are in MB
850+
Int machine_mem = if defined(mem) then mem * 1000 else 9000
851+
Int command_mem = machine_mem - 500
852+
853+
command {
854+
set -e
855+
856+
export GATK_LOCAL_JAR=${default="/root/gatk.jar" gatk_override}
857+
858+
gatk --java-options "-Xmx${command_mem}m" FilterAlignmentArtifacts \
859+
-V ${input_vcf} \
860+
-I ${bam} \
861+
--bwa-mem-index-image ${realignment_index_bundle} \
862+
${realignment_extra_args} \
863+
-O ${output_vcf}
864+
}
865+
866+
runtime {
867+
docker: gatk_docker
868+
bootDiskSizeGb: 12
782869
memory: command_mem + " MB"
783870
disks: "local-disk " + select_first([disk_space, 100]) + if use_ssd then " SSD" else " HDD"
784871
preemptible: select_first([preemptible_attempts, 10])
@@ -842,7 +929,7 @@ task oncotate_m2 {
842929
fi
843930
844931
${default="/root/oncotator_venv/bin/oncotator" oncotator_exe} --db-dir onco_dbdir/ -c $HOME/tx_exact_uniprot_matches.AKT1_CRLF2_FGFR1.txt \
845-
-v ${m2_vcf} ${case_id}.maf.annotated hg19 -i VCF -o TCGAMAF --skip-no-alt --infer-onps --collapse-number-annotations --log_name oncotator.log \
932+
-v ${m2_vcf} ${case_id}.maf.annotated hg19 -i VCF -o TCGAMAF --skip-no-alt --collapse-number-annotations --log_name oncotator.log \
846933
-a Center:${default="Unknown" sequencing_center} \
847934
-a source:${default="Unknown" sequence_source} \
848935
-a normal_barcode:${control_id} \
@@ -906,12 +993,14 @@ task Funcotate {
906993
Array[String]? transcript_selection_list
907994
Array[String]? annotation_defaults
908995
Array[String]? annotation_overrides
996+
Boolean filter_funcotations
909997

910998
# ==============
911999
# Process input args:
9121000
String transcript_selection_arg = if defined(transcript_selection_list) then " --transcript-list " else ""
9131001
String annotation_def_arg = if defined(annotation_defaults) then " --annotation-default " else ""
9141002
String annotation_over_arg = if defined(annotation_overrides) then " --annotation-override " else ""
1003+
String filter_funcotations_args = if (filter_funcotations) then " --remove-filtered-variants " else ""
9151004
# ==============
9161005
9171006
# runtime
@@ -962,11 +1051,13 @@ task Funcotate {
9621051
${"--transcript-selection-mode " + transcript_selection_mode} \
9631052
${transcript_selection_arg}${default="" sep=" --transcript-list " transcript_selection_list} \
9641053
${annotation_def_arg}${default="" sep=" --annotation-default " annotation_defaults} \
965-
${annotation_over_arg}${default="" sep=" --annotation-override " annotation_overrides}
1054+
${annotation_over_arg}${default="" sep=" --annotation-override " annotation_overrides} \
1055+
${filter_funcotations_args}
9661056
>>>
9671057

9681058
runtime {
9691059
docker: gatk_docker
1060+
bootDiskSizeGb: 12
9701061
memory: machine_mem + " MB"
9711062
disks: "local-disk " + select_first([disk_space_gb, default_disk_space_gb]) + if use_ssd then " SSD" else " HDD"
9721063
preemptible: select_first([preemptible_attempts, 3])

0 commit comments

Comments
 (0)