Skip to content

Commit 9029c23

Browse files
authored
Merge pull request #89 from icgc-argo/payload-gen-variant-calling.0.3.3.0
add support for mutect2 callable stats and filtering stats
2 parents bf4dd7c + 452c63e commit 9029c23

7 files changed

Lines changed: 27 additions & 7 deletions
Binary file not shown.
Binary file not shown.
1003 Bytes
Binary file not shown.
377 Bytes
Binary file not shown.

tools/payload-gen-variant-calling/payload-gen-variant-calling.nf

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -22,7 +22,7 @@
2222
*/
2323

2424
nextflow.enable.dsl = 2
25-
version = '0.3.2.0'
25+
version = '0.3.3.0'
2626

2727
params.normal_analysis = ""
2828
params.tumour_analysis = ""

tools/payload-gen-variant-calling/payload-gen-variant-calling.py

Lines changed: 13 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -33,8 +33,8 @@
3333

3434

3535
variant_type_to_data_type_etc = {
36-
'snv': ['Simple Nucleotide Variation', 'Raw SNV Calls', ['CaVEMan'], ['Mutect2']], # dataCategory, dataType, analysis_tools
37-
'indel': ['Simple Nucleotide Variation', 'Raw InDel Calls', ['Pindel'], ['Mutect2']],
36+
'snv': ['Simple Nucleotide Variation', 'Raw SNV Calls', ['CaVEMan'], ['GATK-Mutect2']], # dataCategory, dataType, analysis_tools
37+
'indel': ['Simple Nucleotide Variation', 'Raw InDel Calls', ['Pindel'], ['GATK-Mutect2']],
3838
'cnv': ['Copy Number Variation', 'Raw CNV Calls', ['ASCAT']],
3939
'sv': ['Structural Variation', 'Raw SV Calls', ['BRASS']],
4040
'caveman-supplement': ['Simple Nucleotide Variation', 'Variant Calling Supplement', ['CaVEMan']],
@@ -43,9 +43,11 @@
4343
'brass-supplement': ['Structural Variation', 'Variant Calling Supplement', ['BRASS']],
4444
'timings-supplement': [None, 'Variant Calling Supplement', ['CaVEMan', 'Pindel', 'ASCAT', 'BRASS']],
4545
'bas_metrics': ['Quality Control Metrics', 'Alignment QC', ['bas_stats']],
46-
'contamination_metrics': ['Quality Control Metrics', 'Cross Sample Contamination', ['verifyBamHomChk'], ['CalculateContamination']],
46+
'contamination_metrics': ['Quality Control Metrics', 'Cross Sample Contamination', ['verifyBamHomChk'], ['GATK-CalculateContamination']],
4747
'ascat_metrics': ['Quality Control Metrics', 'Ploidy and Purity Estimation', ['ASCAT']],
4848
'genotyped_gender_metrics': ['Quality Control Metrics', 'Genotyping Inferred Gender', ['compareBamGenotypes']],
49+
'mutect_filtering_metrics': ['Quality Control Metrics', 'Mutect2 Filtering Stats', [], ['GATK-FilterMutectCalls']],
50+
'mutect_callable_metrics': ['Quality Control Metrics', 'Mutect2 Callabe Stats', [], ['GATK-Mutect2']],
4951
}
5052

5153
workflow_full_name = {
@@ -119,9 +121,14 @@ def get_files_info(file_to_upload, wf_short_name, wf_version, somatic_or_germli
119121
elif file_to_upload.endswith('mutect2-indel.vcf.gz') or file_to_upload.endswith('mutect2-indel.vcf.gz.tbi'):
120122
variant_type = 'indel'
121123
elif file_to_upload.endswith('_metrics.tgz'):
122-
variant_type = file_to_upload.split(".")[-2]
123-
if re.match(r'.+?\.normal.contamination_metrics.tgz', file_to_upload):
124-
fname_sample_part = normal_analysis['samples'][0]['sampleId']
124+
if file_to_upload.endswith('.contamination_metrics.tgz'):
125+
variant_type = file_to_upload.split(".")[-2]
126+
if re.match(r'.+?\.normal.contamination_metrics.tgz', file_to_upload):
127+
fname_sample_part = normal_analysis['samples'][0]['sampleId']
128+
elif file_to_upload.endswith('.filtering_metrics.tgz'):
129+
variant_type = 'mutect_filtering_metrics'
130+
elif file_to_upload.endswith('.callable_metrics.tgz'):
131+
variant_type = 'mutect_callable_metrics'
125132
else:
126133
sys.exit('Error: unknown file type "%s"' % file_to_upload)
127134

Lines changed: 13 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,13 @@
1+
{
2+
"normal_analysis": "data/7ad90309-21df-4345-9903-0921df73456b.sequencing_alignment.analysis.json",
3+
"tumour_analysis": "data/f64ae545-11c1-46f7-8ae5-4511c156f7be.sequencing_alignment.analysis.json",
4+
"files_to_upload": [
5+
"data/74fda0ca6f60168dbb34c00f0d93de07.normal.contamination_metrics.tgz",
6+
"data/bcdc11ebe93861a0d447158a84d845ba.tumour.contamination_metrics.tgz",
7+
"data/filter-mutect-calls.filtering-stats.filtering_metrics.tgz",
8+
"data/merged-mutect-stats.stats.callable_metrics.tgz"
9+
],
10+
"wf_name": "gatk-mutect2-variant-calling",
11+
"wf_short_name": "gatk-mutect2",
12+
"wf_version": "0.1.0"
13+
}

0 commit comments

Comments
 (0)