Report on softmask overlaps, and rename softmask stats local module.

oist · Oct 3, 2024 · 2778d59 · 2778d59
1 parent 61f5adc
commit 2778d59
Show file tree

Hide file tree

Showing 3 changed files with 60 additions and 6 deletions.
diff --git a/modules/local/multiqc_softmask_overlaps.nf b/modules/local/multiqc_softmask_overlaps.nf
@@ -0,0 +1,48 @@
+process MULTIQC_SOFTMASK_OVERLAPS {
+    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' }"
+
+    // Recycling bedtools image as we already use it elsewhere
+
+    input:
+    path (files)
+
+    output:
+    path "*_mqc.tsv",  emit: tsv
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    """
+        # Summarise Jaccard indexes for MultiQC
+    echo "# id: 'Jaccard indexes'" > jaccard_indexes_mqc.tsv
+    echo "# section_name: 'Repeat masking overlap statistics'" >> jaccard_indexes_mqc.tsv
+    echo "# format: 'tsv'" >> jaccard_indexes_mqc.tsv
+    echo "# plot_type: 'table'" >> jaccard_indexes_mqc.tsv
+    echo "# description: 'This plot shows a brief summary of the overlaps between the soft masks generated by each process'" >> jaccard_indexes_mqc.tsv
+    echo "# pconfig:" >> jaccard_indexes_mqc.tsv
+    echo "#    id: 'Jaccard indexes'" >> jaccard_indexes_mqc.tsv
+    echo "#    title: 'Jaccard indexes'" >> jaccard_indexes_mqc.tsv
+    echo "#    ylab: ''" >> jaccard_indexes_mqc.tsv
+    echo "id\ttantan–WindowMasker\ttantan–RepeatMasker\tWindowMasker–RepeatMasker" >> jaccard_indexes_mqc.tsv
+    # Here we loop on samples
+    for SAMPLE in \$(basename -s _tantan_windowmasker_jaccard.txt *_tantan_windowmasker_jaccard.txt)
+    do
+        printf "\${SAMPLE}\t"                                                               >> jaccard_indexes_mqc.tsv
+        sed -n 2p \${SAMPLE}_tantan_windowmasker_jaccard.txt       | awk '{printf \$3"\t"}' >> jaccard_indexes_mqc.tsv
+        sed -n 2p \${SAMPLE}_tantan_repeatmasker_jaccard.txt       | awk '{printf \$3"\t"}' >> jaccard_indexes_mqc.tsv
+        sed -n 2p \${SAMPLE}_repeatmasker_windowmasker_jaccard.txt | awk '{printf \$3    }' >> jaccard_indexes_mqc.tsv
+        printf '\n' >> jaccard_indexes_mqc.tsv
+    done
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        bedtools: \$(bedtools --version | sed -e "s/bedtools v//g")
+    END_VERSIONS
+    """
+}
diff --git a/modules/local/multiqc_softmask_statistics.nf b/modules/local/multiqc_softmask_statistics.nf
@@ -1,4 +1,4 @@
-process SOFTMASK_STATS {
+process MULTIQC_SOFTMASK_STATS {
     label 'process_single'
     conda "${moduleDir}/environment.yml"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
@@ -19,8 +19,8 @@ process SOFTMASK_STATS {
     def args = task.ext.args ?: ''
     """
     # Here we make the header
-    echo "# id: 'repeat summary'" > masking_stats_mqc.tsv
-    echo "# section_name: 'repeat masking summary statistics'" >> masking_stats_mqc.tsv
+    echo "# id: 'Repeat summary'" > masking_stats_mqc.tsv
+    echo "# section_name: 'Repeat masking summary statistics'" >> masking_stats_mqc.tsv
     echo "# format: 'tsv'" >> masking_stats_mqc.tsv
     echo "# plot_type: 'table'" >> masking_stats_mqc.tsv
     echo "# description: 'This plot shows a brief summary of each genomes whose repeats has been masked'" >> masking_stats_mqc.tsv

diff --git a/workflows/pairgenomealignmask.nf b/workflows/pairgenomealignmask.nf
@@ -30,7 +30,8 @@ include { SEQTK_CUTN           as EXTLIB_BED                 } from '../modules/
 
 include { MERGE_MASKS          as MERGEDMASKS                } from '../modules/local/mergemasks.nf'
 include { GFASTATS             as MERGEDMASKS_STATS          } from '../modules/nf-core/gfastats/main'
-include { SOFTMASK_STATS                                     } from '../modules/local/multiqc_softmask_statistics.nf'
+include { MULTIQC_SOFTMASK_STATS                             } from '../modules/local/multiqc_softmask_statistics.nf'
+include { MULTIQC_SOFTMASK_OVERLAPS                          } from '../modules/local/multiqc_softmask_overlaps.nf'
 
 include { MULTIQC                     } from '../modules/nf-core/multiqc/main'
 include { paramsSummaryMap            } from 'plugin/nf-validation'
@@ -124,7 +125,7 @@ workflow PAIRGENOMEALIGNMASK {
 
     // Aggregation of statistics
     //
-    SOFTMASK_STATS ( channel.empty()
+    MULTIQC_SOFTMASK_STATS ( channel.empty()
         .mix(        TANTAN_STATS.out.assembly_summary.map {it[1]} )
         .mix(  WINDOWMASKER_STATS.out.assembly_summary.map {it[1]} )
         .mix( REPEATMODELER_STATS.out.assembly_summary.map {it[1]} )
@@ -133,7 +134,12 @@ workflow PAIRGENOMEALIGNMASK {
         .mix(   MERGEDMASKS_STATS.out.assembly_summary.map {it[1]} )
         .collect()
     )
-    ch_multiqc_files = ch_multiqc_files.mix(SOFTMASK_STATS.out.tsv)
+    ch_multiqc_files = ch_multiqc_files.mix(MULTIQC_SOFTMASK_STATS.out.tsv)
+
+    // Aggregation of statistics (Jaccard indices)
+    //
+    MULTIQC_SOFTMASK_OVERLAPS ( MERGEDMASKS.out.txt.map{it[1]}.collect() )
+    ch_multiqc_files = ch_multiqc_files.mix(MULTIQC_SOFTMASK_OVERLAPS.out.tsv)
 
 
     // Collect software versions