Merge pull request #497 from FriederikeHanssen/cnv

Refactor Variant Calling
nf-core · Mar 24, 2022 · 9d74728 · 9d74728
2 parents baf1718 + 38bb8de
commit 9d74728
Show file tree

Hide file tree

Showing 35 changed files with 2,141 additions and 1,013 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -25,25 +25,28 @@ jobs:
         # Nextflow versions
         include:
           # Test pipeline minimum Nextflow version
-          - NXF_VER: '21.10.3'
-            NXF_EDGE: ''
+          - NXF_VER: "21.10.3"
+            NXF_EDGE: ""
           # Test latest edge release of Nextflow
-          - NXF_VER: ''
-            NXF_EDGE: '1'
+          - NXF_VER: ""
+            NXF_EDGE: "1"
         test:
-          - 'aligner'
-          - 'annotation'
-          - 'default'
-          - 'deepvariant'
-          - 'gatk4_spark'
-          - 'haplotypecaller'
-          - 'manta'
+          - "aligner"
+          - "annotation"
+          - "default"
+          - "deepvariant"
+          - "freebayes"
+          - "gatk4_spark"
+          - "haplotypecaller"
+          - "manta"
+          - "mutect2"
+          - "msisensorpro"
           # - 'save_bam_mapped'
-          - 'skip_markduplicates'
-          - 'strelka'
-          - 'split_fastq'
-          - 'targeted'
-          - 'tumor_normal_pair'
+          - "skip_markduplicates"
+          - "strelka"
+          - "split_fastq"
+          - "targeted"
+          - "tumor_normal_pair"
     steps:
       - name: Check out pipeline code
         uses: actions/checkout@v2
@@ -61,7 +64,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v2
         with:
-          python-version: '3.x'
+          python-version: "3.x"
       - name: Install dependencies
         run: python -m pip install --upgrade pip pytest-workflow
 

diff --git a/conf/modules.config b/conf/modules.config
@@ -448,9 +448,6 @@ process{
     }
 
     // DEEPVARIANT
-    withName: 'BGZIP_VC_DEEPVARIANT_GVCF' {
-        ext.when         = { params.generate_gvcf && !params.no_intervals }
-    }
     withName: 'CONCAT_DEEPVARIANT_.*' {
         publishDir       = [
             enabled: "${!params.no_intervals}",
@@ -472,15 +469,7 @@ process{
             pattern: "*{vcf.gz,vcf.gz.tbi}"
         ]
     }
-    withName : 'TABIX_VC_DEEPVARIANT_GVCF' {
-        publishDir       = [
-            enabled: "${params.generate_gvcf}",
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/variant_calling/${meta.id}/deepvariant" },
-            pattern: "*{vcf.gz,vcf.gz.tbi}"
-        ]
-    }
-    withName : 'TABIX_VC_DEEPVARIANT_VCF' {
+    withName : 'TABIX_VC_DEEPVARIANT.*' {
         publishDir       = [
             enabled: true,
             mode: params.publish_dir_mode,
@@ -524,7 +513,7 @@ process{
         ]
     }
     withName: 'HAPLOTYPECALLER' {
-        ext.args         = '-ERC GVCF'
+        ext.args         = { params.joint_germline ? "-ERC GVCF" : "" }
         ext.prefix       = {"${meta.id}.g"}
         ext.when         = { params.tools && params.tools.contains('haplotypecaller') }
         publishDir       = [
@@ -535,7 +524,7 @@ process{
         ]
     }
     withName: 'GENOTYPEGVCFS' {
-        ext.when         = { params.tools && params.tools.contains('haplotypecaller') }
+        ext.when         = { params.tools && params.tools.contains('haplotypecaller') && params.joint_germline}
         publishDir       = [
             enabled: true,
             mode: params.publish_dir_mode,
@@ -616,73 +605,129 @@ process{
 
 // TUMOR_VARIANT_CALLING
 
-    withName: 'MERGEMUTECTSTATS' {
-        ext.prefix       = { "${meta.id}.vcf.gz" }
-    }
-    withName: 'GATHERPILEUPSUMMARIES' {
-        ext.prefix       = { "${meta.id}.table" }
+    //MANTA
+    withName: 'CONCAT_MANTA_TUMOR' {
+        ext.prefix       = {"${meta.id}.tumor_sv"}
     }
 
-// PAIR_VARIANT_CALLING
+    //MUTECT2
+    withName: 'GATK4_CALCULATECONTAMINATION' {
+        publishDir        = [
+            enabled: true,
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
+        ]
+    }
 
-    withName: 'MUTECT2'{
+    withName: 'CONCAT_MUTECT2.*' {
         publishDir       = [
-            enabled: "${params.no_intervals}",
+            enabled: "${!params.no_intervals}",
             mode: params.publish_dir_mode,
-            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
+            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" },
+            pattern: "*{vcf.gz,vcf.gz.tbi}"
         ]
     }
-    withName: 'GATK4_MUTECT2'{
+
+    withName: 'FILTERMUTECTCALLS.*'{
+        ext.prefix       = {"${meta.id}.filtered"}
         publishDir       = [
-            enabled: "${params.no_intervals}",
+            enabled: true,
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
         ]
     }
-    withName: 'CONCAT_MUTECT2' {
+
+    withName: 'GATHERPILEUPSUMMARIES.*' {
+        ext.prefix       = { "${meta.id}.table" }
+        ext.when         = { "${!params.no_intervals}"}
         publishDir       = [
             enabled: "${!params.no_intervals}",
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
         ]
     }
-    withName: 'GATK4_MERGEMUTECTSTATS' {
-        publishDir       = [
-            enabled: true,
+
+    withName: 'GETPILEUPSUMMARIES.*' {
+        publishDir        = [
+            enabled: "${params.no_intervals}",
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
         ]
     }
-    withName: 'GATK4_FILTERMUTECTCALLS'{
-        ext.prefix       = {"${meta.id}.filtered."}
+
+    withName: 'MERGEMUTECTSTATS' {
+        ext.prefix       = { "${meta.id}.vcf.gz" }
         publishDir       = [
             enabled: true,
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" }
         ]
     }
+
+    withName: 'MUTECT2'{
+        ext.when         = { params.tools && params.tools.contains('mutect2') }
+        ext.args         = { params.ignore_soft_clipped_bases ? "--dont-use-soft-clipped-bases true" : "" }
+        publishDir       = [
+            enabled: "${params.no_intervals}",
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" },
+            pattern: "*{gz,gz.tbi,stats}"
+        ]
+    }
+
+// PAIR_VARIANT_CALLING
+
+    //MANTA
+    withName: 'CONCAT_MANTA_SOMATIC' {
+        ext.prefix       = {"${meta.id}.somatic_sv"}
+    }
+
+    //MUTECT2
+    withName: 'CALCULATECONTAMINATION'{
+        //ext.args         = { params.ignore_soft_clipped_bases ? "--dont-use-soft-clipped-bases true" : "" }
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" },
+        ]
+    }
+
+    withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:GATK_TUMOR_NORMAL_SOMATIC_VARIANT_CALLING:GATHERPILEUPSUMMARIES.*' {
+        ext.prefix       = { "${meta.id}.table" }
+        publishDir       = [
+            enabled: "${!params.no_intervals}",
+            mode: params.publish_dir_mode,
+            //use ${meta.tumor_id}_vs_${meta_normal_id} to publish in the same directory as the remainders of the
+            //somatic output whilst keeping the filename prefix identifieable for status type
+            path: { "${params.outdir}/variant_calling/${meta.tumor_id}_vs_${meta.normal_id}/mutect2" }
+        ]
+    }
+
+    withName: 'LEARNREADORIENTATIONMODEL'{
+        ext.prefix         = { "${meta.id}.learnreadorientationmodel" }
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/${meta.id}/mutect2" },
+        ]
+    }
+
+    //MSISENSORPRO
+    withName: 'MSISENSORPRO_MSI_SOMATIC'{
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/variant_calling/${meta.id}/msisensorpro" },
+        ]
+    }
+
+    //STRELKA
+    withName: 'CONCAT_STRELKA_INDELS' {
+        ext.prefix       = {"${meta.id}.somatic_indels"}
+    }
+    withName: 'CONCAT_STRELKA_SNVS' {
+        ext.prefix       = {"${meta.id}.somatic_snvs"}
+    }
+
 }
-// withName: 'GATK4_CALCULATECONTAMINATION'{
-//    ext.args          = ''
-//     publishDir        = [
-//         enabled: false,
-//         mode: params.publish_dir_mode
-//     ]
-//}
-//withName: 'GATK4_FILTERMUTECTCALLS'{
-//    ext.args          = ''
-//     publishDir        = [
-//         enabled: false,
-//         mode: params.publish_dir_mode
-//     ]
-//}
-//withName: 'GATK4_GETPILEUPSUMMARIES'{
-//    ext.args          = ''
-//     publishDir        = [
-//         enabled: false,
-//         mode: params.publish_dir_mode
-//     ]
-//}
+
 //withName: 'GENOMICSDBIMPORT' {
 //
 //}

diff --git a/conf/test.config b/conf/test.config
@@ -91,6 +91,30 @@ profiles {
         params.input               = "${baseDir}/tests/csv/3.0/recalibrated_germline.csv"
         params.dbsnp               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz"
         params.fasta               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta"
+        params.intervals           = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/multi_intervals.bed"
+        params.step                = 'variant_calling'
+        params.joint_germline      = true
+        params.wes                 = true
+        params.genome              = 'WBcel235'
+        params.vep_genome          = 'WBcel235'
+    }
+    tools_tumoronly {
+        params.input               = "${baseDir}/tests/csv/3.0/recalibrated_tumoronly.csv"
+        params.dbsnp               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz"
+        params.fasta               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta"
+        params.germline_resource   = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/gnomAD.r2.1.1.vcf.gz"
+        params.intervals           = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/multi_intervals.bed"
+        params.pon                 = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz"
+        params.step                = 'variant_calling'
+        params.joint_germline      = true
+        params.wes                 = true
+        params.genome              = 'WBcel235'
+        params.vep_genome          = 'WBcel235'
+    }
+    tools_somatic {
+        params.input               = "${baseDir}/tests/csv/3.0/recalibrated_somatic.csv"
+        params.dbsnp               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/dbsnp_138.hg38.vcf.gz"
+        params.fasta               = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/genome.fasta"
         params.germline_resource   = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/gnomAD.r2.1.1.vcf.gz"
         params.intervals           = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/sequence/multi_intervals.bed"
         params.pon                 = "${params.genomes_base}/data/genomics/homo_sapiens/genome/chr21/germlineresources/mills_and_1000G.indels.hg38.vcf.gz"

diff --git a/modules.json b/modules.json
@@ -3,6 +3,9 @@
     "homePage": "https://github.com/nf-core/sarek",
     "repos": {
         "nf-core/modules": {
+            "ascat": {
+                "git_sha": "d6244b42f596fa26d2ecba4ce862755821ed9da8"
+            },
             "bcftools/stats": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
@@ -24,6 +27,9 @@
             "cnvkit/batch": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
+            "controlfreec": {
+                "git_sha": "c189835b1bb444e5ee87416fdbea66e2c2ba365e"
+            },
             "custom/dumpsoftwareversions": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
@@ -108,6 +114,15 @@
             "gatk4/variantrecalibrator": {
                 "git_sha": "e745e167c1020928ef20ea1397b6b4d230681b4d"
             },
+            "manta/germline": {
+                "git_sha": "979e57b7ac6a405a395dd7a6dbe1a275c5bc226b"
+            },
+            "manta/somatic": {
+                "git_sha": "979e57b7ac6a405a395dd7a6dbe1a275c5bc226b"
+            },
+            "manta/tumoronly": {
+                "git_sha": "979e57b7ac6a405a395dd7a6dbe1a275c5bc226b"
+            },
             "msisensorpro/msi_somatic": {
                 "git_sha": "c8ebd0de36c649a14fc92f2f73cbd9f691a8ce0a"
             },