Add indices and meta

This commit is contained in:
Rike 2022-05-17 23:38:26 +02:00
parent 15fba1dd7c
commit 5ae757121f
3 changed files with 73 additions and 36 deletions

View file

@ -8,15 +8,16 @@ process GATK4_FILTERVARIANTTRANCHES {
'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }" 'quay.io/biocontainers/gatk4:4.2.6.1--hdfd78af_0' }"
input: input:
tuple val(meta), path(vcf), path(intervals) tuple val(meta), path(vcf), path(tbi),path(intervals)
path ressources path resources
path resources_index
path fasta path fasta
path fai path fai
path dict path dict
output: output:
tuple val(meta), path("*.bam"), emit: bam tuple val(meta), path("*.vcf.gz"), emit: vcf
path "versions.yml" , emit: versions path "versions.yml" , emit: versions
when: when:
@ -26,7 +27,7 @@ process GATK4_FILTERVARIANTTRANCHES {
def args = task.ext.args ?: '' def args = task.ext.args ?: ''
def prefix = task.ext.prefix ?: "${meta.id}" def prefix = task.ext.prefix ?: "${meta.id}"
def ressources = ressources.collect{"--ressources $it"}.join(' ') def resources = resources.collect{"--resource $it"}.join(' ')
def avail_mem = 3 def avail_mem = 3
if (!task.memory) { if (!task.memory) {
log.info '[GATK FilterVariantTranches] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.' log.info '[GATK FilterVariantTranches] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
@ -36,7 +37,7 @@ process GATK4_FILTERVARIANTTRANCHES {
""" """
gatk --java-options "-Xmx${avail_mem}g" FilterVariantTranches \\ gatk --java-options "-Xmx${avail_mem}g" FilterVariantTranches \\
--variant $vcf \\ --variant $vcf \\
$ressources \\ $resources \\
--output ${prefix}.filtered.vcf.gz \\ --output ${prefix}.filtered.vcf.gz \\
--tmp-dir . \\ --tmp-dir . \\
$args $args

View file

@ -1,51 +1,64 @@
name: "gatk4_filtervarianttranches" name: "gatk4_filtervarianttranches"
## TODO nf-core: Add a description of the module and list keywords description: Apply tranche filtering
description: write your description here
keywords: keywords:
- sort - gatk4
- filtervarianttranches
tools: tools:
- "gatk4": - "gatk4":
## TODO nf-core: Add a description and other details for the software below description: Genome Analysis Toolkit (GATK4)
description: "Genome Analysis Toolkit (GATK4)" homepage: https://gatk.broadinstitute.org/hc/en-us
homepage: "None" documentation: https://gatk.broadinstitute.org/hc/en-us
documentation: "None" tool_dev_url: https://github.com/broadinstitute/gatk
tool_dev_url: "https://github.com/broadinstitute/gatk" doi: "10.1158/1538-7445.AM2017-3590"
doi: "" licence: ["BSD-3-clause"]
licence: "['BSD-3-clause']"
## TODO nf-core: Add a description of all of the variables used as input
input: input:
# Only when we have meta
- meta: - meta:
type: map type: map
description: | description: |
Groovy Map containing sample information Groovy Map containing sample information
e.g. [ id:'test', single_end:false ] e.g. [ id:'test', single_end:false ]
# - vcf:
## TODO nf-core: Delete / customise this example input
- bam:
type: file type: file
description: BAM/CRAM/SAM file description: a VCF file containing variants, must have info key:CNN_2D
pattern: "*.{bam,cram,sam}" 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:
## TODO nf-core: Add a description of all of the variables used as output
output: output:
#Only when we have meta
- meta: - meta:
type: map type: map
description: | description: |
Groovy Map containing sample information Groovy Map containing sample information
e.g. [ id:'test', single_end:false ] e.g. [ id:'test', single_end:false ]
#
- versions: - versions:
type: file type: file
description: File containing software versions description: File containing software versions
pattern: "versions.yml" pattern: "versions.yml"
## TODO nf-core: Delete / customise this example output - vcf:
- bam:
type: file type: file
description: Sorted BAM/CRAM/SAM file description: VCF file
pattern: "*.{bam,cram,sam}" pattern: "*.vcf.gz"
authors: authors:
- "@FriederikeHanssen" - "@FriederikeHanssen"

View file

@ -6,10 +6,33 @@ include { GATK4_FILTERVARIANTTRANCHES } from '../../../../modules/gatk4/filterva
workflow test_gatk4_filtervarianttranches { workflow test_gatk4_filtervarianttranches {
input = [ input = [ [ id:'test' ], // meta map
[ id:'test', single_end:false ], // meta map file(params.test_data['homo_sapiens']['illumina']['test2_haplotc_ann_vcf_gz'], checkIfExists: true),
file(params.test_data['sarscov2']['illumina']['test_paired_end_bam'], checkIfExists: true) file(params.test_data['homo_sapiens']['illumina']['test2_haplotc_ann_vcf_gz_tbi'], checkIfExists: true),
[]
] ]
GATK4_FILTERVARIANTTRANCHES ( input ) resources = [
file(params.test_data['homo_sapiens']['genome']['hapmap_3_3_hg38_21_vcf_gz'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['res_1000g_omni2_5_hg38_21_vcf_gz'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['res_1000g_phase1_snps_hg38_21_vcf_gz'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['dbsnp_138_hg38_21_vcf_gz'], checkIfExists: true)
]
resources_index =[ file(params.test_data['homo_sapiens']['genome']['hapmap_3_3_hg38_21_vcf_gz_tbi'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['res_1000g_omni2_5_hg38_21_vcf_gz_tbi'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['res_1000g_phase1_snps_hg38_21_vcf_gz_tbi'], checkIfExists: true),
//file(params.test_data['homo_sapiens']['genome']['dbsnp_138_hg38_21_vcf_gz_tbi'], checkIfExists: true)
]
// , [
// 'hapmap,known=false,training=true,truth=true,prior=15.0 hapmap_3.3.hg38.vcf.gz',
// 'omni,known=false,training=true,truth=false,prior=12.0 1000G_omni2.5.hg38.vcf.gz',
// '1000G,known=false,training=true,truth=false,prior=10.0 1000G_phase1.snps.hg38.vcf.gz',
// 'dbsnp,known=true,training=false,truth=false,prior=2.0 dbsnp_138.hg38.vcf.gz'
// ]]
fasta = file(params.test_data['homo_sapiens']['genome']['genome_21_fasta'], checkIfExists: true)
fai = file(params.test_data['homo_sapiens']['genome']['genome_21_fasta_fai'], checkIfExists: true)
dict = file(params.test_data['homo_sapiens']['genome']['genome_21_dict'], checkIfExists: true)
GATK4_FILTERVARIANTTRANCHES ( input, resources, resources_index, fasta, fai, dict)
} }