adding homer merge peaks, and a peakcalling workflow

modules/nf-core/homer/annotatepeaks/environment.yml

@@ -4,9 +4,9 @@ channels:
   - conda-forge
   - bioconda
 dependencies:
-  - bioconda::bioconductor-deseq2=1.42.0
-  - bioconda::bioconductor-edger=4.0.16
-  - bioconda::homer=4.11
-  - bioconda::samtools=1.21
-  - conda-forge::r-essentials=4.3
+  - bioconda::bioconductor-deseq2=1.46.0
+  - bioconda::bioconductor-edger=4.4.0
+  - bioconda::homer=5.1
+  - bioconda::samtools=1.22.1
+  - conda-forge::r-essentials=4.4
   - conda-forge::wget=1.21.4

modules/nf-core/homer/annotatepeaks/main.nf

@@ -1,10 +1,13 @@
 process HOMER_ANNOTATEPEAKS {
     tag "${meta.id}"
     label 'process_medium'
+
     conda "${moduleDir}/environment.yml"
+    // singularity build url: https://wave.seqera.io/view/builds/bd-9c603739ae7d4fd3_1
+    // docker build url: https://wave.seqera.io/view/builds/bd-08c7bb832e96c6bd_1
     container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
-        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/0f/0fe4a3875b78dce3c66b43fb96489769cc32e55e329e2525d2af09096af2252a/data'
-        : 'community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:a8f4c58755bb281b'}"
+        ? 'oras://community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:9c603739ae7d4fd3'
+        : 'community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:08c7bb832e96c6bd'}"
 
     input:
     tuple val(meta), path(peak)
@@ -22,7 +25,7 @@ process HOMER_ANNOTATEPEAKS {
     script:
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def VERSION = '4.11'
+    def VERSION = '5.1'
     // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
     """
     annotatePeaks.pl \\
@@ -41,7 +44,7 @@ process HOMER_ANNOTATEPEAKS {
 
     stub:
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def VERSION = '4.11'
+    def VERSION = '5.1'
     // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
     """
     touch ${prefix}.annotatePeaks.txt

modules/nf-core/homer/findpeaks/environment.yml

@@ -4,9 +4,9 @@ channels:
   - conda-forge
   - bioconda
 dependencies:
-  - bioconda::bioconductor-deseq2=1.42.0
-  - bioconda::bioconductor-edger=4.0.16
-  - bioconda::homer=4.11
-  - bioconda::samtools=1.21
-  - conda-forge::r-essentials=4.3
+  - bioconda::bioconductor-deseq2=1.46.0
+  - bioconda::bioconductor-edger=4.4.0
+  - bioconda::homer=5.1
+  - bioconda::samtools=1.22.1
+  - conda-forge::r-essentials=4.4
   - conda-forge::wget=1.21.4

modules/nf-core/homer/findpeaks/main.nf

@@ -1,34 +1,65 @@
 process HOMER_FINDPEAKS {
     tag "${meta.id}"
     label 'process_medium'
+
     conda "${moduleDir}/environment.yml"
+    // singularity build url: https://wave.seqera.io/view/builds/bd-9c603739ae7d4fd3_1
+    // docker build url: https://wave.seqera.io/view/builds/bd-08c7bb832e96c6bd_1
     container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
-        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/0f/0fe4a3875b78dce3c66b43fb96489769cc32e55e329e2525d2af09096af2252a/data'
-        : 'community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:a8f4c58755bb281b'}"
+        ? 'oras://community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:9c603739ae7d4fd3'
+        : 'community.wave.seqera.io/library/bioconductor-deseq2_bioconductor-edger_homer_samtools_pruned:08c7bb832e96c6bd'}"
 
     input:
+    val(style)
     tuple val(meta), path(tagDir)
+    tuple val(meta2), path(controlTagDir)
     path uniqmap
 
     output:
-    tuple val(meta), path("*.peaks.txt"), emit: txt
+    tuple val(meta), path("*.txt"), emit: txt
+    tuple val(meta), path("*.gtf"), optional: true, emit: gtf
     path "versions.yml", emit: versions
 
     when:
     task.ext.when == null || task.ext.when
 
     script:
     def args = task.ext.args ?: ''
-    def prefix = task.ext.prefix ?: uniqmap ? "${meta.id}-${uniqmap.baseName}" : "${meta.id}"
+    def prefix = task.ext.prefix ?: "${meta.id}"
     def uniqmap_flag = uniqmap ? "-uniqmap ${uniqmap}" : ""
-    def VERSION = '4.11'
-    // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
-    """
+    def control_flag = controlTagDir ? "-i ${controlTagDir}" : ""
+    def control_suffix = controlTagDir ? "_vs_${meta2.id}" : ""
+    def VERSION = '5.1'
+
+    // Validate style
+    def valid_styles = ['factor', 'histone', 'groseq', 'tss', 'dnase', 'super', 'mC']
+    if (!valid_styles.contains(style)) {
+        error "style must be one of: ${valid_styles.join(', ')}"
+    }
 
+    // Determine output suffix based on style
+    def output_suffix_map = [
+        'factor': 'peaks.txt',
+        'histone': 'regions.txt',
+        'groseq': 'transcripts.txt',
+        'tss': 'tss.txt',
+        'dnase': 'peaks.txt',
+        'super': 'superEnhancers.txt',
+        'mC': 'regions.txt'
+    ]
+    def output_suffix = output_suffix_map[style]
+    def output_filename = "${prefix}${control_suffix}_${output_suffix}"
+
+    def groseq_gtf_arg = style == "groseq" ? "-gtf ${prefix}${control_suffix}_transcripts.gtf" : ""
+
+    """
     findPeaks \\
         ${tagDir} \\
+        -style ${style} \\
+        ${control_flag} \\
         ${args} \\
-        -o ${prefix}.peaks.txt \\
+        ${groseq_gtf_arg} \\
+        -o ${output_filename} \\
         ${uniqmap_flag}
 
     cat <<-END_VERSIONS > versions.yml
@@ -39,11 +70,25 @@ process HOMER_FINDPEAKS {
 
     stub:
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def VERSION = '4.11'
-    // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
+    def control_suffix = controlTagDir ? "_vs_${meta2.id}" : ""
+    def VERSION = '5.1'
+
+    def output_suffix_map = [
+        'factor': 'peaks.txt',
+        'histone': 'regions.txt',
+        'groseq': 'transcripts.txt',
+        'tss': 'tss.txt',
+        'dnase': 'peaks.txt',
+        'super': 'superEnhancers.txt',
+        'mC': 'regions.txt'
+    ]
+    def output_suffix = output_suffix_map[style]
+    def output_filename = "${prefix}${control_suffix}_${output_suffix}"
+    def gtf_filename = "${prefix}${control_suffix}_transcripts.gtf"
 
     """
-    touch ${prefix}.peaks.txt
+    touch ${output_filename}
+    ${style == 'groseq' ? "touch ${gtf_filename}" : ''}
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

modules/nf-core/homer/findpeaks/meta.yml

@@ -15,6 +15,14 @@ tools:
       licence: ["GPL-3.0-or-later"]
       identifier: biotools:homer
 input:
+  - style:
+      type: string
+      description: |
+        One of factor, histone, groseq, tss, dnase, super or mC. See the homer
+        findPeak documentation for more information. The choice of style does
+        affect which parameters must be passed in through `ext.args`, and affects
+        the suffix of the output file (e.g. if style is factor, then the file output
+        to the txt channel will be *_peaks.txt. See output for more information.)
   - - meta:
         type: map
         description: |
@@ -24,10 +32,19 @@ input:
         type: directory
         description: "The 'Tag Directory'"
         pattern: "tagDir"
+  - - meta2:
+        type: map
+        description: |
+          Groovy map containing sample information for the control sample
+          e.g. [ id: "control", single_end: false ]
+    - controlTagDir:
+        type: directory
+        description: "The 'Tag Directory' for the control sample"
+        pattern: "controlTagDir"
   - uniqmap:
       type: directory
       description: "(directory of binary files specifying uniquely mappable locations)
-        Download from http://biowhat.ucsd.edu/homer/groseq/"
+        Download from http://biowha t.ucsd.edu/homer/groseq/"
       pattern: "uniqmap/"
 output:
   txt:
@@ -36,7 +53,21 @@ output:
           description: |
             Groovy Map containing sample information
             e.g. [ id:'test', single_end:false ]
-      - "*.peaks.txt":
+      - "*.txt":
+          type: file
+          description: |
+            One of *_peaks.txt, *_regions, *_transcripts.txt,
+            *_tss.txt, *_superEnhancer.txt. Which depends on the `style`, e.g. for
+            style factor, the output txt file would be *_peaks.txt
+          pattern: "*.txt"
+          ontologies: []
+  gtf:
+    - - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+      - "*.gtf":
           type: file
           description: "Peaks in txt file"
           pattern: "*.peaks.txt"
@@ -52,3 +83,4 @@ authors:
   - "@edmundmiller"
 maintainers:
   - "@edmundmiller"
+  - "@cmatkhan"

modules/nf-core/homer/findpeaks/tests/main.nf.test

@@ -49,8 +49,10 @@ nextflow_process {
         when {
             process {
                 """
-                input[0] = HOMER_MAKETAGDIRECTORY.out.tagdir
-                input[1] = []
+                input[0] = "factor"
+                input[1] = HOMER_MAKETAGDIRECTORY.out.tagdir
+                input[2] = [[:], []]
+                input[3] = []
                 """
             }
         }
@@ -71,8 +73,10 @@ nextflow_process {
         when {
             process {
                 """
-                input[0] = HOMER_MAKETAGDIRECTORY.out.tagdir
-                input[1] = []
+                input[0] = "factor"
+                input[1] = HOMER_MAKETAGDIRECTORY.out.tagdir
+                input[2] = [[:], []]
+                input[3] = []
                 """
             }
         }
@@ -92,8 +96,10 @@ nextflow_process {
         when {
             process {
                 """
-                input[0] = HOMER_MAKETAGDIRECTORY.out.tagdir
-                input[1] = UNZIP.out.unzipped_archive.map { it[1] }
+                input[0] = "factor"
+                input[1] = HOMER_MAKETAGDIRECTORY.out.tagdir
+                input[2] = [[:], []]
+                input[3] = UNZIP.out.unzipped_archive.map { it[1] }
                 """
             }
         }
@@ -111,4 +117,4 @@ nextflow_process {
             )
         }
     }
-}
+}