From 38c559b2a8b9a8cda62fa215521f581eb979c773 Mon Sep 17 00:00:00 2001
From: Ramprasad Neethiraj <20065894+ramprasadn@users.noreply.github.com>
Date: Wed, 9 Aug 2023 13:15:52 +0000
Subject: [PATCH] Add GATK's CreateReadCountPanelOfNormals and
 DenoiseReadCounts (#3709)

* pon

* denoisereadcounts

* add stub

* update version string

* review suggestions

* Update modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf [skip ci]

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>

* Update modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf [skip ci]

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>

* Update modules/nf-core/gatk4/denoisereadcounts/main.nf [skip ci]

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>

* Update modules/nf-core/gatk4/denoisereadcounts/main.nf

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>

---------

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>
---
 .../createreadcountpanelofnormals/main.nf     | 54 +++++++++++++++++
 .../createreadcountpanelofnormals/meta.yml    | 46 +++++++++++++++
 .../nf-core/gatk4/denoisereadcounts/main.nf   | 59 +++++++++++++++++++
 .../nf-core/gatk4/denoisereadcounts/meta.yml  | 59 +++++++++++++++++++
 tests/config/pytest_modules.yml               |  8 +++
 .../createreadcountpanelofnormals/main.nf     | 37 ++++++++++++
 .../nextflow.config                           | 13 ++++
 .../createreadcountpanelofnormals/test.yml    | 17 ++++++
 .../nf-core/gatk4/denoisereadcounts/main.nf   | 40 +++++++++++++
 .../gatk4/denoisereadcounts/nextflow.config   | 13 ++++
 .../nf-core/gatk4/denoisereadcounts/test.yml  | 19 ++++++
 11 files changed, 365 insertions(+)
 create mode 100644 modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
 create mode 100644 modules/nf-core/gatk4/createreadcountpanelofnormals/meta.yml
 create mode 100644 modules/nf-core/gatk4/denoisereadcounts/main.nf
 create mode 100644 modules/nf-core/gatk4/denoisereadcounts/meta.yml
 create mode 100644 tests/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
 create mode 100644 tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config
 create mode 100644 tests/modules/nf-core/gatk4/createreadcountpanelofnormals/test.yml
 create mode 100644 tests/modules/nf-core/gatk4/denoisereadcounts/main.nf
 create mode 100644 tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config
 create mode 100644 tests/modules/nf-core/gatk4/denoisereadcounts/test.yml

diff --git a/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf b/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
new file mode 100644
index 00000000000..1f30dceee39
--- /dev/null
+++ b/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
@@ -0,0 +1,54 @@
+process GATK4_CREATEREADCOUNTPANELOFNORMALS {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "bioconda::gatk4=4.4.0.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gatk4:4.4.0.0--py36hdfd78af_0':
+        'biocontainers/gatk4:4.4.0.0--py36hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(counts)
+
+    output:
+    tuple val(meta), path("*.hdf5"), emit: pon
+    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 input_list = counts.collect(){"--input $it"}.join(" ")
+
+    def avail_mem = 3072
+    if (!task.memory) {
+        log.info '[GATK CreateReadCountPanelOfNormals] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = (task.memory.mega*0.8).intValue()
+    }
+    """
+    gatk --java-options "-Xmx${avail_mem}M" CreateReadCountPanelOfNormals \\
+        ${args} \\
+        ${input_list} \\
+        --output ${prefix}.hdf5
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.hdf5
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/gatk4/createreadcountpanelofnormals/meta.yml b/modules/nf-core/gatk4/createreadcountpanelofnormals/meta.yml
new file mode 100644
index 00000000000..f86ac768952
--- /dev/null
+++ b/modules/nf-core/gatk4/createreadcountpanelofnormals/meta.yml
@@ -0,0 +1,46 @@
+name: "gatk4_createreadcountpanelofnormals"
+description: Creates a panel of normals (PoN) for read-count denoising given the read counts for samples in the panel.
+keywords:
+  - gatk4
+  - createreadcountpanelofnormals
+  - panelofnormals
+tools:
+  - gatk4:
+      description: |
+        Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+      doi: 10.1158/1538-7445.AM2017-3590
+      tool_dev_url: "https://github.com/broadinstitute/gatk"
+      licence: ["Apache-2.0"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - counts:
+      type: file
+      description: Read counts in hdf5 or tsv format.
+      pattern: "*.{hdf5,tsv}"
+
+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"
+  - pon:
+      type: file
+      description: Panel-of-normals file.
+      pattern: "*.{hdf5}"
+
+authors:
+  - "@ramprasadn"
diff --git a/modules/nf-core/gatk4/denoisereadcounts/main.nf b/modules/nf-core/gatk4/denoisereadcounts/main.nf
new file mode 100644
index 00000000000..e60aeda1925
--- /dev/null
+++ b/modules/nf-core/gatk4/denoisereadcounts/main.nf
@@ -0,0 +1,59 @@
+process GATK4_DENOISEREADCOUNTS {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "bioconda::gatk4=4.4.0.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gatk4:4.4.0.0--py36hdfd78af_0':
+        'biocontainers/gatk4:4.4.0.0--py36hdfd78af_0' }"
+
+    input:
+    tuple val(meta), path(counts)
+    tuple val(meta2), path(pon)
+
+    output:
+    tuple val(meta), path("*_standardizedCR.tsv"), emit: standardized
+    tuple val(meta), path("*_denoisedCR.tsv")    , emit: denoised
+    path "versions.yml"                          , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    def avail_mem = 3072
+    if (!task.memory) {
+        log.info '[GATK DenoiseReadCounts] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this.'
+    } else {
+        avail_mem = (task.memory.mega*0.8).intValue()
+    }
+    """
+    gatk --java-options "-Xmx${avail_mem}M" DenoiseReadCounts \\
+        ${args} \\
+        --tmp-dir . \\
+        --input ${counts} \\
+        --count-panel-of-normals ${pon} \\
+        --standardized-copy-ratios ${prefix}_standardizedCR.tsv \\
+        --denoised-copy-ratios ${prefix}_denoisedCR.tsv
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}_standardizedCR.tsv
+    touch ${prefix}_denoisedCR.tsv
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gatk4: \$(echo \$(gatk --version 2>&1) | sed 's/^.*(GATK) v//; s/ .*\$//')
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/gatk4/denoisereadcounts/meta.yml b/modules/nf-core/gatk4/denoisereadcounts/meta.yml
new file mode 100644
index 00000000000..0bb2f389d00
--- /dev/null
+++ b/modules/nf-core/gatk4/denoisereadcounts/meta.yml
@@ -0,0 +1,59 @@
+name: "gatk4_denoisereadcounts"
+description: Denoises read counts to produce denoised copy ratios
+keywords:
+  - gatk4
+  - denoisereadcounts
+  - copyratios
+tools:
+  - gatk4:
+      description: |
+        Developed in the Data Sciences Platform at the Broad Institute, the toolkit offers a wide variety of tools
+        with a primary focus on variant discovery and genotyping. Its powerful processing engine
+        and high-performance computing features make it capable of taking on projects of any size.
+      homepage: https://gatk.broadinstitute.org/hc/en-us
+      documentation: https://gatk.broadinstitute.org/hc/en-us/categories/360002369672s
+      doi: 10.1158/1538-7445.AM2017-3590
+      tool_dev_url: "https://github.com/broadinstitute/gatk"
+      licence: ["Apache-2.0"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - meta2:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - counts:
+      type: file
+      description: Read counts in hdf5 or tsv format.
+      pattern: "*.{hdf5,tsv}"
+  - pon:
+      type: file
+      description: Panel of normals file hdf5 or tsv format.
+      pattern: "*.{hdf5}"
+
+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"
+  - standardized:
+      type: file
+      description: Standardized copy ratios file.
+      pattern: "*.{tsv}"
+  - denoised:
+      type: file
+      description: Denoised copy ratios file
+      pattern: "*.{tsv}"
+
+authors:
+  - "@ramprasadn"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 0144bc892ef..b7cf7e17e6f 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -1363,6 +1363,10 @@ gatk4/condensedepthevidence:
   - modules/nf-core/gatk4/condensedepthevidence/**
   - tests/modules/nf-core/gatk4/condensedepthevidence/**
 
+gatk4/createreadcountpanelofnormals:
+  - modules/nf-core/gatk4/createreadcountpanelofnormals/**
+  - tests/modules/nf-core/gatk4/createreadcountpanelofnormals/**
+
 gatk4/createsequencedictionary:
   - modules/nf-core/gatk4/createsequencedictionary/**
   - tests/modules/nf-core/gatk4/createsequencedictionary/**
@@ -1371,6 +1375,10 @@ gatk4/createsomaticpanelofnormals:
   - modules/nf-core/gatk4/createsomaticpanelofnormals/**
   - tests/modules/nf-core/gatk4/createsomaticpanelofnormals/**
 
+gatk4/denoisereadcounts:
+  - modules/nf-core/gatk4/denoisereadcounts/**
+  - tests/modules/nf-core/gatk4/denoisereadcounts/**
+
 gatk4/determinegermlinecontigploidy:
   - modules/nf-core/gatk4/determinegermlinecontigploidy/**
   - tests/modules/nf-core/gatk4/determinegermlinecontigploidy/**
diff --git a/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
new file mode 100644
index 00000000000..f45144db259
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf
@@ -0,0 +1,37 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_CREATEREADCOUNTPANELOFNORMALS } from '../../../../../modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf'
+include { GATK4_COLLECTREADCOUNTS             } from '../../../../../modules/nf-core/gatk4/collectreadcounts/main.nf'
+include { GATK4_PREPROCESSINTERVALS           } from '../../../../../modules/nf-core/gatk4/preprocessintervals/main.nf'
+
+workflow test_gatk4_createreadcountpanelofnormals {
+
+    fasta = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)]).collect()
+    fai   = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)]).collect()
+    dict  = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)]).collect()
+
+    GATK4_PREPROCESSINTERVALS ( fasta, fai, dict, [[],[]], [[],[]]).interval_list
+        .map {meta,list -> list}
+        .set {ch_intervals}
+
+    input = Channel.of([
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true),
+    ],
+    [
+        [ id:'test2', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam_bai'], checkIfExists: true),
+    ]).combine(ch_intervals)
+
+    GATK4_COLLECTREADCOUNTS ( input, fasta, fai, dict )
+
+    GATK4_CREATEREADCOUNTPANELOFNORMALS (
+        GATK4_COLLECTREADCOUNTS.out.tsv
+            .map({ meta, tsv -> [ [id:'test'], tsv ] })
+            .groupTuple()
+    )
+}
diff --git a/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config
new file mode 100644
index 00000000000..a8c5250b024
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config
@@ -0,0 +1,13 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: GATK4_COLLECTREADCOUNTS {
+        ext.args = "--format TSV --interval-merging-rule OVERLAPPING_ONLY"
+    }
+
+    withName: GATK4_CREATEREADCOUNTPANELOFNORMALS {
+        ext.args = "--minimum-interval-median-percentile 1.0 --number-of-eigensamples 2"
+    }
+
+}
diff --git a/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/test.yml b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/test.yml
new file mode 100644
index 00000000000..829079feaa5
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/createreadcountpanelofnormals/test.yml
@@ -0,0 +1,17 @@
+- name: "gatk4 createreadcountpanelofnormals"
+  command: nextflow run ./tests/modules/nf-core/gatk4/createreadcountpanelofnormals -entry test_gatk4_createreadcountpanelofnormals -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config
+  tags:
+    - "gatk4"
+    - "gatk4/createreadcountpanelofnormals"
+  files:
+    - path: "output/gatk4/test.hdf5"
+    - path: "output/gatk4/versions.yml"
+
+- name: "gatk4 createreadcountpanelofnormals stub"
+  command: nextflow run ./tests/modules/nf-core/gatk4/createreadcountpanelofnormals -entry test_gatk4_createreadcountpanelofnormals -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/gatk4/createreadcountpanelofnormals/nextflow.config -stub
+  tags:
+    - "gatk4"
+    - "gatk4/createreadcountpanelofnormals"
+  files:
+    - path: "output/gatk4/test.hdf5"
+    - path: "output/gatk4/versions.yml"
diff --git a/tests/modules/nf-core/gatk4/denoisereadcounts/main.nf b/tests/modules/nf-core/gatk4/denoisereadcounts/main.nf
new file mode 100644
index 00000000000..e2bd1b69401
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/denoisereadcounts/main.nf
@@ -0,0 +1,40 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { GATK4_CREATEREADCOUNTPANELOFNORMALS } from '../../../../../modules/nf-core/gatk4/createreadcountpanelofnormals/main.nf'
+include { GATK4_COLLECTREADCOUNTS             } from '../../../../../modules/nf-core/gatk4/collectreadcounts/main.nf'
+include { GATK4_PREPROCESSINTERVALS           } from '../../../../../modules/nf-core/gatk4/preprocessintervals/main.nf'
+include { GATK4_DENOISEREADCOUNTS             } from '../../../../../modules/nf-core/gatk4/denoisereadcounts/main.nf'
+
+workflow test_gatk4_denoisereadcounts {
+
+    fasta = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)]).collect()
+    fai   = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)]).collect()
+    dict  = Channel.of([ [ id:'test' ], file(params.test_data['homo_sapiens']['genome']['genome_dict'], checkIfExists: true)]).collect()
+
+    GATK4_PREPROCESSINTERVALS ( fasta, fai, dict, [[],[]], [[],[]]).interval_list
+        .map {meta,list -> list}
+        .set {ch_intervals}
+
+    input = Channel.of([
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true),
+    ],
+    [
+        [ id:'test2', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test2_paired_end_sorted_bam_bai'], checkIfExists: true),
+    ])
+
+    GATK4_COLLECTREADCOUNTS ( input.combine(ch_intervals), fasta, fai, dict )
+
+    GATK4_CREATEREADCOUNTPANELOFNORMALS (
+        GATK4_COLLECTREADCOUNTS.out.tsv
+            .map({ meta, tsv -> [ [id:'test'], tsv ] })
+            .groupTuple()
+    )
+
+    GATK4_DENOISEREADCOUNTS ( GATK4_COLLECTREADCOUNTS.out.tsv.first(), GATK4_CREATEREADCOUNTPANELOFNORMALS.out.pon )
+}
diff --git a/tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config b/tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config
new file mode 100644
index 00000000000..a8c5250b024
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config
@@ -0,0 +1,13 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: GATK4_COLLECTREADCOUNTS {
+        ext.args = "--format TSV --interval-merging-rule OVERLAPPING_ONLY"
+    }
+
+    withName: GATK4_CREATEREADCOUNTPANELOFNORMALS {
+        ext.args = "--minimum-interval-median-percentile 1.0 --number-of-eigensamples 2"
+    }
+
+}
diff --git a/tests/modules/nf-core/gatk4/denoisereadcounts/test.yml b/tests/modules/nf-core/gatk4/denoisereadcounts/test.yml
new file mode 100644
index 00000000000..c0fd4a85b7e
--- /dev/null
+++ b/tests/modules/nf-core/gatk4/denoisereadcounts/test.yml
@@ -0,0 +1,19 @@
+- name: "gatk4 denoisereadcounts"
+  command: nextflow run ./tests/modules/nf-core/gatk4/denoisereadcounts -entry test_gatk4_denoisereadcounts -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config
+  tags:
+    - "gatk4"
+    - "gatk4/denoisereadcounts"
+  files:
+    - path: "output/gatk4/test_standardizedCR.tsv"
+    - path: "output/gatk4/test_denoisedCR.tsv"
+    - path: "output/gatk4/versions.yml"
+
+- name: "gatk4 denoisereadcounts stub"
+  command: nextflow run ./tests/modules/nf-core/gatk4/denoisereadcounts -entry test_gatk4_denoisereadcounts -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/gatk4/denoisereadcounts/nextflow.config -stub
+  tags:
+    - "gatk4"
+    - "gatk4/denoisereadcounts"
+  files:
+    - path: "output/gatk4/test_standardizedCR.tsv"
+    - path: "output/gatk4/test_denoisedCR.tsv"
+    - path: "output/gatk4/versions.yml"