Merge pull request #1676 from FriederikeHanssen/gatk_filter

Gatk filtervarianttranches

modules/gatk4/cnnscorevariants/main.nf

@@ -9,7 +9,7 @@ process GATK4_CNNSCOREVARIANTS {
     container 'broadinstitute/gatk:4.2.6.1' //Biocontainers is missing a package
 
     input:
-    tuple val(meta), path(vcf), path(aligned_input), path(intervals)
+    tuple val(meta), path(vcf), path(tbi), path(aligned_input), path(intervals)
     path fasta
     path fai
     path dict
@@ -17,8 +17,9 @@ process GATK4_CNNSCOREVARIANTS {
     path weights
 
     output:
-    tuple val(meta), path("*.vcf.gz"), emit: vcf
+    tuple val(meta), path("*cnn.vcf.gz")    , emit: vcf
-    path "versions.yml"           , emit: versions
+    tuple val(meta), path("*cnn.vcf.gz.tbi"), emit: tbi
+    path "versions.yml"                     , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
@@ -40,7 +41,7 @@ process GATK4_CNNSCOREVARIANTS {
     """
     gatk --java-options "-Xmx${avail_mem}g" CNNScoreVariants \\
         --variant $vcf \\
-        --output ${prefix}.vcf.gz \\
+        --output ${prefix}.cnn.vcf.gz \\
         --reference $fasta \\
         $interval_command \\
         $aligned_input \\

modules/gatk4/cnnscorevariants/meta.yml

@@ -25,6 +25,10 @@ input:
       type: file
       description: VCF file
       pattern: "*.vcf.gz"
+  - tbi:
+      type: file
+      description: VCF index file
+      pattern: "*.vcf.gz.tbi"
   - aligned_input:
       type: file
       description: BAM/CRAM file from alignment (optional)
@@ -67,6 +71,10 @@ output:
       type: file
       description: Annotated VCF file
       pattern: "*.vcf"
+  - tbi:
+      type: file
+      description: VCF index file
+      pattern: "*.vcf.gz.tbi"
 
 authors:
   - "@FriederikeHanssen"

modules/gatk4/filtervarianttranches/main.nf

@@ -0,0 +1,51 @@
+process GATK4_FILTERVARIANTTRANCHES {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::gatk4=4.2.6.1" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gatk4:4.2.6.1--hdfd78af_0':
+        'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(vcf), path(tbi), path(intervals)
+    path resources
+    path resources_index
+    path fasta
+    path fai
+    path dict
+
+
+    output:
+    tuple val(meta), path("*.vcf.gz")    , emit: vcf
+    tuple val(meta), path("*.vcf.gz.tbi"), emit: tbi
+    path "versions.yml"                  , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    def resources = resources.collect{"--resource $it"}.join(' ')
+    def avail_mem = 3
+    if (!task.memory) {
+        log.info '[GATK FilterVariantTranches] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = task.memory.giga
+    }
+    """
+    gatk --java-options "-Xmx${avail_mem}g" FilterVariantTranches \\
+        --variant $vcf \\
+        $resources \\
+        --output ${prefix}.filtered.vcf.gz \\
+        --tmp-dir . \\
+        $args
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}

modules/gatk4/filtervarianttranches/meta.yml

@@ -0,0 +1,68 @@
+name: "gatk4_filtervarianttranches"
+description: Apply tranche filtering
+keywords:
+  - gatk4
+  - filtervarianttranches
+
+tools:
+  - "gatk4":
+      description: Genome Analysis Toolkit (GATK4)
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us
+      tool_dev_url: https://github.com/broadinstitute/gatk
+      doi: "10.1158/1538-7445.AM2017-3590"
+      licence: ["BSD-3-clause"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - vcf:
+      type: file
+      description: a VCF file containing variants, must have info key:CNN_2D
+      pattern: "*.vcf.gz"
+  - tbi:
+      type: file
+      description: tbi file matching with -vcf
+      pattern: "*.vcf.gz.tbi"
+  - resources:
+      type: list
+      description: resource A VCF containing known SNP and or INDEL sites. Can be supplied as many times as necessary
+      pattern: "*.vcf.gz"
+  - resources_index:
+      type: list
+      description: Index of resource VCF containing known SNP and or INDEL sites. Can be supplied as many times as necessary
+      pattern: "*.vcf.gz"
+  - fasta:
+      type: file
+      description: The reference fasta file
+      pattern: "*.fasta"
+  - fai:
+      type: file
+      description: Index of reference fasta file
+      pattern: "fasta.fai"
+  - dict:
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - vcf:
+      type: file
+      description: VCF file
+      pattern: "*.vcf.gz"
+  - tbi:
+      type: file
+      description: VCF index file
+      pattern: "*.vcf.gz.tbi"
+
+authors:
+  - "@FriederikeHanssen"

tests/config/pytest_modules.yml

@@ -759,6 +759,10 @@ gatk4/filtermutectcalls:
   - modules/gatk4/filtermutectcalls/**
   - tests/modules/gatk4/filtermutectcalls/**
 
+gatk4/filtervarianttranches:
+  - modules/gatk4/filtervarianttranches/**
+  - tests/modules/gatk4/filtervarianttranches/**
+
 gatk4/gatherbqsrreports:
   - modules/gatk4/gatherbqsrreports/**
   - tests/modules/gatk4/gatherbqsrreports/**

tests/config/test_data.config

@@ -266,6 +266,8 @@ params {
 
                 test2_haplotc_ann_vcf_gz                       = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test2_haplotc.ann.vcf.gz"
                 test2_haplotc_ann_vcf_gz_tbi                   = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test2_haplotc.ann.vcf.gz.tbi"
+                test_haplotc_cnn_vcf_gz                        = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test_haplotcaller.cnn.vcf.gz"
+                test_haplotc_cnn_vcf_gz_tbi                    = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test_haplotcaller.cnn.vcf.gz.tbi"
 
                 test2_haplotc_vcf_gz                           = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test2_haplotc.vcf.gz"
                 test2_haplotc_vcf_gz_tbi                       = "${test_data_dir}/genomics/homo_sapiens/illumina/gatk/haplotypecaller_calls/test2_haplotc.vcf.gz.tbi"

tests/modules/gatk4/cnnscorevariants/main.nf

@@ -7,7 +7,8 @@ include { GATK4_CNNSCOREVARIANTS } from '../../../../modules/gatk4/cnnscorevaria
 workflow test_gatk4_cnnscorevariants {
 
     input     = [ [ id:'test' ], // meta map
-                    file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf'], checkIfExists: true),
+                    file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf_gz'], checkIfExists: true),
+                    file(params.test_data['homo_sapiens']['illumina']['test_genome_vcf_gz_tbi'], checkIfExists: true),
                     [],
                     []
                 ]

tests/modules/gatk4/cnnscorevariants/test.yml

@@ -1,9 +1,9 @@
 - name: gatk4 cnnscorevariants test_gatk4_cnnscorevariants
   command: nextflow run ./tests/modules/gatk4/cnnscorevariants -entry test_gatk4_cnnscorevariants -c ./tests/config/nextflow.config  -c ./tests/modules/gatk4/cnnscorevariants/nextflow.config
   tags:
-    - gatk4
     - gatk4/cnnscorevariants
+    - gatk4
   files:
-    - path: output/gatk4/test.vcf.gz
+    - path: output/gatk4/test.cnn.vcf.gz
-      contains:
+      contains: ["##ALT=<ID=NON_REF,Description="]
-        - "##ALT=<ID=NON_REF,Description="
+    - path: output/gatk4/test.cnn.vcf.gz.tbi

tests/modules/gatk4/filtervarianttranches/main.nf

@@ -0,0 +1,26 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_FILTERVARIANTTRANCHES } from '../../../../modules/gatk4/filtervarianttranches/main.nf'
+include { GATK4_CNNSCOREVARIANTS } from '../../../../modules/gatk4/cnnscorevariants/main.nf'
+include { GATK4_HAPLOTYPECALLER  } from '../../../../modules/gatk4/haplotypecaller/main.nf'
+workflow test_gatk4_filtervarianttranches {
+
+    resources = [ file(params.test_data['homo_sapiens']['genome']['dbsnp_146_hg38_vcf_gz'], checkIfExists: true) ]
+    resources_index  = [
+                    file(params.test_data['homo_sapiens']['genome']['dbsnp_146_hg38_vcf_gz_tbi'], checkIfExists: true),
+                ]
+
+    input = [ [ id:'test' ], // meta map
+                file(params.test_data['homo_sapiens']['illumina']['test_haplotc_cnn_vcf_gz'], checkIfExists: true),
+                file(params.test_data['homo_sapiens']['illumina']['test_haplotc_cnn_vcf_gz_tbi'], checkIfExists: true),
+                []
+            ]
+
+    fasta  = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fai    = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+    dict   = file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)
+
+    GATK4_FILTERVARIANTTRANCHES (input , resources, resources_index, fasta, fai, dict)
+}

tests/modules/gatk4/filtervarianttranches/nextflow.config

@@ -0,0 +1,6 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    ext.args = "--info-key CNN_1D"
+}

tests/modules/gatk4/filtervarianttranches/test.yml

@@ -0,0 +1,9 @@
+- name: gatk4 filtervarianttranches test_gatk4_filtervarianttranches
+  command: nextflow run ./tests/modules/gatk4/filtervarianttranches -entry test_gatk4_filtervarianttranches -c ./tests/config/nextflow.config  -c ./tests/modules/gatk4/filtervarianttranches/nextflow.config
+  tags:
+    - gatk4/filtervarianttranches
+    - gatk4
+  files:
+    - path: output/gatk4/test.filtered.vcf.gz
+    - path: output/gatk4/test.filtered.vcf.gz.tbi
+    - path: output/gatk4/versions.yml