Update GOTC single-sample workflow to functional equivalence version

scripts/broad_pipelines/PublicPairedSingleSampleWf_170412.inputs.json

@@ -0,0 +1,113 @@
+{
+  "##_COMMENT1": "SAMPLE NAME AND UNMAPPED BAMS",
+  "PairedEndSingleSampleWorkflow.sample_name": "NA12878",
+  "PairedEndSingleSampleWorkflow.base_file_name": "NA12878_20k",
+  "PairedEndSingleSampleWorkflow.flowcell_unmapped_bams": [
+    "gs://genomics-public-data/test-data/dna/wgs/hiseq2500/NA12878/H06HDADXX130110.1.ATCACGAT.20k_reads.bam",
+    "gs://genomics-public-data/test-data/dna/wgs/hiseq2500/NA12878/H06HDADXX130110.2.ATCACGAT.20k_reads.bam",
+    "gs://genomics-public-data/test-data/dna/wgs/hiseq2500/NA12878/H06JUADXX130110.1.ATCACGAT.20k_reads.bam"
+  ],
+  "PairedEndSingleSampleWorkflow.final_gvcf_name": "NA12878.g.vcf.gz",
+  "PairedEndSingleSampleWorkflow.unmapped_bam_suffix": ".bam",
+  
+  "##_COMMENT2": "INTERVALS",  
+  "PairedEndSingleSampleWorkflow.scattered_calling_intervals": [
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0001_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0002_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0003_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0004_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0005_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0006_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0007_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0008_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0009_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0010_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0011_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0012_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0013_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0014_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0015_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0016_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0017_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0018_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0019_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0020_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0021_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0022_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0023_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0024_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0025_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0026_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0027_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0028_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0029_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0030_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0031_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0032_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0033_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0034_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0035_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0036_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0037_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0038_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0039_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0040_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0041_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0042_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0043_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0044_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0045_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0046_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0047_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0048_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0049_of_50/scattered.interval_list", 
+    "gs://genomics-public-data/resources/broad/hg38/v0/scattered_calling_intervals/temp_0050_of_50/scattered.interval_list"
+  ],
+  "PairedEndSingleSampleWorkflow.wgs_calling_interval_list": "gs://genomics-public-data/resources/broad/hg38/v0/wgs_calling_regions.hg38.interval_list",
+  
+  "##_COMMENT2": "REFERENCE FILES", 
+  "PairedEndSingleSampleWorkflow.ref_dict": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.dict",
+  "PairedEndSingleSampleWorkflow.ref_fasta": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta",
+  "PairedEndSingleSampleWorkflow.ref_fasta_index": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.fai",
+  "PairedEndSingleSampleWorkflow.ref_alt": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.alt",
+  "PairedEndSingleSampleWorkflow.ref_sa": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.sa",
+  "PairedEndSingleSampleWorkflow.ref_amb": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.amb",
+  "PairedEndSingleSampleWorkflow.ref_bwt": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.bwt",
+  "PairedEndSingleSampleWorkflow.ref_ann": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.ann",
+  "PairedEndSingleSampleWorkflow.ref_pac": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.fasta.64.pac",
+  
+  "##_COMMENT3": "KNOWN SITES RESOURCES", 
+  "PairedEndSingleSampleWorkflow.dbSNP_vcf": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.dbsnp138.vcf",
+  "PairedEndSingleSampleWorkflow.dbSNP_vcf_index": "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.dbsnp138.vcf.idx",
+  "PairedEndSingleSampleWorkflow.known_indels_sites_VCFs": [
+    "gs://genomics-public-data/resources/broad/hg38/v0/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz",
+    "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz"
+  ],
+  "PairedEndSingleSampleWorkflow.known_indels_sites_indices": [
+    "gs://genomics-public-data/resources/broad/hg38/v0/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz.tbi",
+    "gs://genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi"
+  ],
+
+  "##_COMMENT4": "QUALITY CONTROL RESOURCES",
+  "PairedEndSingleSampleWorkflow.contamination_sites_vcf": "gs://gatk-test-data/qc/WholeGenomeShotgunContam.vcf",
+  "PairedEndSingleSampleWorkflow.contamination_sites_vcf_index": "gs://gatk-test-data/qc/WholeGenomeShotgunContam.vcf.idx",
+  "PairedEndSingleSampleWorkflow.haplotype_database_file": "gs://gatk-test-data/qc/empty.haplotype_map.txt",
+  "PairedEndSingleSampleWorkflow.fingerprint_genotypes_file": "gs://gatk-test-data/qc/empty.fingerprint.vcf",  
+  "PairedEndSingleSampleWorkflow.wgs_coverage_interval_list": "gs://gatk-test-data/intervals/wgs_coverage_regions.hg38.interval_list",
+  "PairedEndSingleSampleWorkflow.wgs_evaluation_interval_list": "gs://gatk-test-data/intervals/wgs_evaluation_regions.hg38.interval_list",
+
+  "##_COMMENT5": "QUALITY CONTROL SETTINGS (to override defaults)",
+  "PairedEndSingleSampleWorkflow.ValidateReadGroupSamFile.ignore": ["null"],
+  "PairedEndSingleSampleWorkflow.ValidateReadGroupSamFile.max_output": 1000000000,
+  "PairedEndSingleSampleWorkflow.ValidateAggregatedSamFile.ignore": ["null"],
+  "PairedEndSingleSampleWorkflow.ValidateAggregatedSamFile.max_output": 1000000000,
+
+  "##_COMMENT6": "DISK SIZES + PREEMPTIBLES",
+  "PairedEndSingleSampleWorkflow.agg_small_disk": 200,
+  "PairedEndSingleSampleWorkflow.agg_medium_disk": 300,
+  "PairedEndSingleSampleWorkflow.agg_large_disk": 400,
+  "PairedEndSingleSampleWorkflow.agg_preemptible_tries": 3,
+  "PairedEndSingleSampleWorkflow.flowcell_small_disk": 100,
+  "PairedEndSingleSampleWorkflow.flowcell_medium_disk": 200,
+  "PairedEndSingleSampleWorkflow.preemptible_tries": 3
+}

scripts/broad_pipelines/PublicPairedSingleSampleWf_170412.options.json

@@ -0,0 +1,3 @@
+{
+	"read_from_cache":false
+}