Collect statistics from the all-maskers-combined genome

oist · Oct 3, 2024 · 9217db1 · 9217db1
1 parent 3ca8049
commit 9217db1
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 4 deletions.
diff --git a/conf/modules.config b/conf/modules.config
@@ -25,6 +25,10 @@ process {
         time       = { check_max( 30.h   * task.attempt, 'time'    ) }
     }
 
+    withName: 'MERGEDMASKS_STATS' {
+        ext.prefix = { "${meta.id}_allmaskers" }
+    }
+
     withName: 'MULTIQC' {
         ext.args   = { params.multiqc_title ? "--title \"$params.multiqc_title\"" : '' }
         publishDir = [

diff --git a/modules/local/multiqc_softmask_statistics.nf b/modules/local/multiqc_softmask_statistics.nf
@@ -28,7 +28,7 @@ process SOFTMASK_STATS {
     echo "#    id: 'repeat summary'" >> masking_stats_mqc.tsv
     echo "#    title: 'repeat summary'" >> masking_stats_mqc.tsv
     echo "#    ylab: ''" >> masking_stats_mqc.tsv
-    echo "id\tTotal scaffold length\tTotal contig length\ttantan masked bases\twindowmasker masked bases\trmodeler masked bases (REPM)\trmodeler masked bases (DFAM)\trmodeler masked bases (EXTR)" >> masking_stats_mqc.tsv
+    echo "id\tTotal scaffold length\tTotal contig length\ttantan masked bases\twindowmasker masked bases\trmodeler masked bases (REPM)\trmodeler masked bases (DFAM)\trmodeler masked bases (EXTR)\tAll maskers combined" >> masking_stats_mqc.tsv
     # Here we loop on samples
     for file in *_tantan.assembly_summary
     do
@@ -47,6 +47,7 @@ process SOFTMASK_STATS {
                 printf '\tNA' >> masking_stats_mqc.tsv
             fi
         done
+        grep 'soft-masked bases' \${SAMPLE}_allmaskers.assembly_summary | awk '{printf "\t"\$4}' >> masking_stats_mqc.tsv
         printf '\n' >> masking_stats_mqc.tsv
     done
     """

diff --git a/workflows/pairgenomealignmask.nf b/workflows/pairgenomealignmask.nf
@@ -29,6 +29,7 @@ include { GFASTATS             as EXTLIB_STATS               } from '../modules/
 include { SEQTK_CUTN           as EXTLIB_BED                 } from '../modules/local/seqtk.nf'
 
 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                     } from '../modules/nf-core/multiqc/main'
@@ -115,10 +116,11 @@ workflow PAIRGENOMEALIGNMASK {
     //
     MERGEDMASKS (
         input_genomes
-            .join(TANTAN_BED.out.bed_gz.map{meta, bed -> [ [id:meta.key ] , bed ] } )
-            .join(WINDOWMASKER_BED.out.bed_gz.map{meta, bed -> [ [id:meta.key ] , bed ] })
-            .join(REPEATMODELER_BED.out.bed_gz.map{meta, bed -> [ [id:meta.key ] , bed ] })
+            .join(TANTAN_BED.out.bed_gz.map       {meta, bed -> [ [id:meta.key ] , bed ] } )
+            .join(WINDOWMASKER_BED.out.bed_gz.map {meta, bed -> [ [id:meta.key ] , bed ] } )
+            .join(REPEATMODELER_BED.out.bed_gz.map{meta, bed -> [ [id:meta.key ] , bed ] } )
     )
+    MERGEDMASKS_STATS ( MERGEDMASKS.out.fasta )
 
     // Aggregation of statistics
     //
@@ -128,6 +130,7 @@ workflow PAIRGENOMEALIGNMASK {
         .mix( REPEATMODELER_STATS.out.assembly_summary.map {it[1]} )
         .mix(          DFAM_STATS_maybeout            .map {it[1]} )
         .mix(        EXTLIB_STATS_maybeout            .map {it[1]} )
+        .mix(   MERGEDMASKS_STATS.out.assembly_summary.map {it[1]} )
         .collect()
     )
     ch_multiqc_files = ch_multiqc_files.mix(SOFTMASK_STATS.out.tsv)