Merge pull request #1 from charles-plessy/compareMasks

Compare and combine masked regions

conf/modules.config

@@ -54,6 +54,19 @@ process {
     withName: 'WINDOWMASKER_USTAT' {
         ext.args = { "-outfmt fasta" }
     }
+
+    withName: 'REPEATMODELER_BED' {
+        ext.prefix = { "${meta.id}.repeatmasker" }
+    }
+
+    withName: 'TANTAN_BED' {
+        ext.prefix = { "${meta.id}.tantan" }
+    }
+
+    withName: 'WINDOWMASKER_BED' {
+        ext.prefix = { "${meta.id}.windowmasker" }
+    }
+
     withName: 'MULTIQC' {
         ext.args   = { params.multiqc_title ? "--title \"$params.multiqc_title\"" : '' }
         publishDir = [

modules/local/bedtools.nf

@@ -0,0 +1,48 @@
+process BEDTOOLS_CUSTOM {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/bedtools:2.31.1--hf5e1c6e_0' :
+        'biocontainers/bedtools:2.31.1--hf5e1c6e_0' }"
+
+    input:
+    tuple val(meta), path(genome), path(tantan), path(windowmasker), path(repeatmasker)
+
+    output:
+    tuple val(meta), path("*.fasta.gz")    , emit: fasta
+    tuple val(meta), path("*_jaccard.txt") , emit: txt
+    tuple val(meta), path("*.bed.gz")      , emit: bed_gz
+    path "versions.yml"                    , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    bedtools jaccard -nonamecheck -a $tantan       -b $windowmasker         > ${prefix}_tantan_windowmasker_jaccard.txt
+    bedtools jaccard -nonamecheck -a $tantan       -b $repeatmasker         > ${prefix}_tantan_repeatmasker_jaccard.txt
+    bedtools jaccard -nonamecheck -a $repeatmasker -b $windowmasker         > ${prefix}_repeatmasker_windowmasker_jaccard.txt
+    
+    zcat $tantan $windowmasker               | sort -k1,1 -k2,2n | bedtools merge | gzip --best  > ${prefix}_tantan_windowmasker.bed.gz
+    zcat $tantan $repeatmasker               | sort -k1,1 -k2,2n | bedtools merge | gzip --best  > ${prefix}_tantan_repeatmasker.bed.gz
+    zcat $windowmasker $repeatmasker         | sort -k1,1 -k2,2n | bedtools merge | gzip --best  > ${prefix}_windowmasker_repeatmasker.bed.gz
+    zcat $tantan $windowmasker $repeatmasker | sort -k1,1 -k2,2n | bedtools merge | gzip --best  > ${prefix}_allmaskers.bed.gz
+
+    bedtools \\
+        maskfasta \\
+        -soft \\
+        -fi $genome \\
+        -bed ${prefix}_allmaskers.bed.gz \\
+        -fo /dev/stdout |
+        gzip --best > ${prefix}_allmaskers.fasta.gz
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        bedtools: \$(bedtools --version | sed -e "s/bedtools v//g")
+    END_VERSIONS
+    """
+}

modules/local/seqtk.nf

@@ -0,0 +1,49 @@
+process SEQTK_CUTN {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/seqtk:1.4--he4a0461_1' :
+        'biocontainers/seqtk:1.4--he4a0461_1' }"
+
+    input:
+    tuple val(meta), path(fasta)
+
+    output:
+    tuple val(meta), path("*.bed.gz") , emit: bed_gz
+    path "versions.yml"               , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    # Produces a compressed BED-3 file with the coordinates of soft-masked regions.
+    awk '/^>/ {print; next} {gsub(/[acgt]/, "N"); print}' $fasta |
+        seqtk cutN -gn 1 - |
+        sort -k1,1 -k2,2n | gzip --best > ${prefix}.mask.bed.gz
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        seqtk: \$(echo \$(seqtk 2>&1) | sed 's/^.*Version: //; s/ .*\$//')
+    END_VERSIONS
+    """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    touch ${prefix}.mask.bed
+    gzip --best ${prefix}.mask.bed
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        seqtk: \$(echo \$(seqtk 2>&1) | sed 's/^.*Version: //; s/ .*\$//')
+    END_VERSIONS
+    """
+
+}

workflows/pairgenomealignmask.nf

@@ -10,7 +10,11 @@ include { REPEATMODELER_REPEATMODELER } from '../modules/nf-core/repeatmodeler/r
 include { REPEATMODELER_MASKER        } from '../modules/nf-core/repeatmodeler/repeatmasker/main'
 include { REPEATMODELER_BUILDDATABASE } from '../modules/nf-core/repeatmodeler/builddatabase/main'
 include { TANTAN                      } from '../modules/local/tantan.nf'
+include { BEDTOOLS_CUSTOM             } from '../modules/local/bedtools.nf'
 include { CUSTOMMODULE                } from '../modules/local/custommodule.nf'
+include { SEQTK_CUTN as TANTAN_BED        } from '../modules/local/seqtk.nf'
+include { SEQTK_CUTN as WINDOWMASKER_BED  } from '../modules/local/seqtk.nf'
+include { SEQTK_CUTN as REPEATMODELER_BED } from '../modules/local/seqtk.nf'
 include { GFASTATS as GFSTTANTAN      } from '../modules/nf-core/gfastats/main'
 include { GFASTATS as GFSTREPEATMOD   } from '../modules/nf-core/gfastats/main'
 include { GFASTATS as GFSTWINDOWMASK  } from '../modules/nf-core/gfastats/main'
@@ -49,7 +53,14 @@ workflow PAIRGENOMEALIGNMASK {
     GFSTTANTAN (
         TANTAN.out.masked_fa
     )
-
+
+    //
+    // MODULE: tantan_bed
+    //
+    TANTAN_BED {
+        TANTAN.out.masked_fa
+    }
+
     // MODULE: repeatmodeler_builddatabase
     //
     REPEATMODELER_BUILDDATABASE (
@@ -77,6 +88,13 @@ workflow PAIRGENOMEALIGNMASK {
         REPEATMODELER_MASKER.out.fasta
     )
 
+    //
+    // MODULE: repeatmodeler_bed
+    //
+    REPEATMODELER_BED {
+        REPEATMODELER_MASKER.out.fasta
+    }
+
     //
     // MODULE: windowmasker_mkcounts
     //
@@ -98,6 +116,13 @@ workflow PAIRGENOMEALIGNMASK {
         WINDOWMASKER_USTAT.out.intervals
     )
 
+    //
+    // MODULE: windowmasker_bed
+    //
+    WINDOWMASKER_BED {
+        WINDOWMASKER_USTAT.out.intervals
+    }
+
     //
     // MODULE: CUSTOMMODULE
     //
@@ -108,8 +133,23 @@ workflow PAIRGENOMEALIGNMASK {
     )
     ch_multiqc_files = ch_multiqc_files.mix(CUSTOMMODULE.out.tsv)
 
+    //
+    // MODULE: bedtools_custom
+    //
+    BEDTOOLS_CUSTOM (
+        ch_samplesheet
+            .join(TANTAN_BED.out.bed_gz)
+            .join(WINDOWMASKER_BED.out.bed_gz)
+            .join(REPEATMODELER_BED.out.bed_gz)
+    )
 
-    ch_versions = ch_versions.mix(WINDOWMASKER_MKCOUNTS.out.versions.first())
+    ch_versions = ch_versions
+        .mix(WINDOWMASKER_MKCOUNTS.out.versions.first())
+        .mix(TANTAN.out.versions.first())
+        .mix(REPEATMODELER_REPEATMODELER.out.versions.first())
+        .mix(GFSTWINDOWMASK.out.versions.first())
+        .mix(TANTAN_BED.out.versions.first())
+        .mix(BEDTOOLS_CUSTOM.out.versions.first())
 
     //
     // Collate and save software versions