[DRAFT] Feature/ploidy support

assets/schema_input.json

@@ -33,6 +33,13 @@
                 "default": 0,
                 "enum": [0, 1]
             },
+            "ploidy": {
+                "type": "integer",
+                "errorMessage": "Ploidy must be a positive integer. Defaults to 2 if none is supplied.",
+                "meta": ["ploidy"],
+                "default": 2,
+                "minimum": 1
+            },
             "lane": {
                 "anyOf": [{ "type": "integer" }, { "type": "string" }],
                 "pattern": "^\\S+$",

conf/modules/haplotypecaller.config

@@ -16,7 +16,9 @@
 process {
 
     withName: 'GATK4_HAPLOTYPECALLER' {
+        // Default ploidy is 2 as most common model organisms (human, mouse, etc.) are diploid
         ext.args   = {  [
+            "--sample-ploidy ${meta.ploidy ?: 2}",
             params.joint_germline          ? "-ERC GVCF"                                          : "",
             params.gatk_pcr_indel_model    ? "--pcr-indel-model ${params.gatk_pcr_indel_model}"   : "",
         ].join(" ").trim() }

conf/modules/modules.config

@@ -114,7 +114,11 @@ process {
 
     withName: 'VCFTOOLS_.*' {
         ext.prefix = { variant_file.baseName - ".vcf" }
-        ext.when   = { !(params.skip_tools && params.skip_tools.split(',').contains('vcftools')) }
+        // VCFtools TsTv does not support polyploid samples (ploidy > 2)
+        ext.when   = {
+            !(params.skip_tools && params.skip_tools.split(',').contains('vcftools')) &&
+            (meta.ploidy == null || meta.ploidy <= 2)
+        }
         publishDir = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/reports/vcftools/${meta.variantcaller}/${meta.id}/" },

conf/test/tools_germline_haploid.config

@@ -0,0 +1,33 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Nextflow config file for running minimal tests with haploid samples
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Defines input files and everything required to run a fast and simple pipeline test.
+    Tests the ploidy parameter with haploid (ploidy=1) samples.
+
+    Use as follows:
+        nextflow run nf-core/sarek -profile test_tools_germline_haploid,<docker/singularity> --outdir <OUTDIR>
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+*/
+
+params {
+    input                  = "${projectDir}/tests/csv/3.0/recalibrated_germline_haploid.csv"
+    genome                 = null
+    igenomes_ignore        = true
+    dbsnp                  = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz"
+    dbsnp_tbi              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz.tbi"
+    fasta                  = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta"
+    fasta_fai              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta.fai"
+    intervals              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/multi_intervals.bed"
+    known_indels           = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz"
+    known_indels_tbi       = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz.tbi"
+    known_indels_vqsr      = "--resource:1000G,known=false,training=true,truth=true,prior=10.0 mills_and_1000G.indels.hg38.vcf.gz"
+    known_snps             = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/hapmap_3.3.hg38.vcf.gz"
+    known_snps_tbi         = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/hapmap_3.3.hg38.vcf.gz.tbi"
+    known_snps_vqsr        = "--resource:hapmap,known=false,training=true,truth=true,prior=10.0 hapmap_3.3.hg38.vcf.gz"
+    ngscheckmate_bed       = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/SNP_GRCh38_hg38_wChr.bed"
+    nucleotides_per_second = 20
+    step                   = 'variant_calling'
+    tools                  = null
+    wes                    = true
+}

conf/test/tools_germline_triploid.config

@@ -0,0 +1,34 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Nextflow config file for running minimal tests with triploid samples
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Defines input files and everything required to run a fast and simple pipeline test.
+    Tests the ploidy parameter with triploid (ploidy=3) samples.
+    VCFtools should be skipped for ploidy > 2.
+
+    Use as follows:
+        nextflow run nf-core/sarek -profile test,tools_germline_triploid,docker --outdir <OUTDIR>
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+*/
+
+params {
+    input                  = "${projectDir}/tests/csv/3.0/recalibrated_germline_triploid.csv"
+    genome                 = null
+    igenomes_ignore        = true
+    dbsnp                  = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz"
+    dbsnp_tbi              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz.tbi"
+    fasta                  = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta"
+    fasta_fai              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta.fai"
+    intervals              = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/sequence/multi_intervals.bed"
+    known_indels           = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz"
+    known_indels_tbi       = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz.tbi"
+    known_indels_vqsr      = "--resource:1000G,known=false,training=true,truth=true,prior=10.0 mills_and_1000G.indels.hg38.vcf.gz"
+    known_snps             = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/hapmap_3.3.hg38.vcf.gz"
+    known_snps_tbi         = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/hapmap_3.3.hg38.vcf.gz.tbi"
+    known_snps_vqsr        = "--resource:hapmap,known=false,training=true,truth=true,prior=10.0 hapmap_3.3.hg38.vcf.gz"
+    ngscheckmate_bed       = "${params.modules_testdata_base_path}/genomics/homo_sapiens/genome/chr21/germlineresources/SNP_GRCh38_hg38_wChr.bed"
+    nucleotides_per_second = 20
+    step                   = 'variant_calling'
+    tools                  = null
+    wes                    = true
+}

nextflow.config

@@ -330,6 +330,8 @@ profiles {
     targeted                   { includeConfig 'conf/test/targeted.config' }
     tools                      { includeConfig 'conf/test/tools.config' }
     tools_germline             { includeConfig 'conf/test/tools_germline.config' }
+    tools_germline_haploid     { includeConfig 'conf/test/tools_germline_haploid.config' }
+    tools_germline_triploid    { includeConfig 'conf/test/tools_germline_triploid.config' }
     tools_somatic              { includeConfig 'conf/test/tools_somatic.config' }
     tools_somatic_ascat        { includeConfig 'conf/test/tools_somatic_ascat.config' }
     tools_tumoronly            { includeConfig 'conf/test/tools_tumoronly.config' }

tests/csv/3.0/recalibrated_germline_haploid.csv

@@ -0,0 +1,2 @@
+patient,sex,status,sample,ploidy,cram,crai
+test,XX,0,sample1,1,https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/illumina/cram/test.paired_end.recalibrated.sorted.cram,https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/illumina/cram/test.paired_end.recalibrated.sorted.cram.crai

tests/csv/3.0/recalibrated_germline_triploid.csv

@@ -0,0 +1,2 @@
+patient,sex,status,sample,ploidy,cram,crai
+test,XX,0,sample1,3,https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/illumina/cram/test.paired_end.recalibrated.sorted.cram,https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/illumina/cram/test.paired_end.recalibrated.sorted.cram.crai