config processing fixed

ghga-de · May 17, 2024 · 46fc4d7 · 46fc4d7
1 parent 2e79eee
commit 46fc4d7
Show file tree

Hide file tree

Showing 3 changed files with 52 additions and 36 deletions.
diff --git a/modules/local/get_contigs.nf b/modules/local/get_contigs.nf
@@ -8,49 +8,64 @@ process GET_CONTIGS {
 
     input:
     tuple val(meta), path(tumor), path(tumor_bai), path(control),  path(control_bai)
+    each path(contig_file)
 
     output:
-    path "contigs.bed"   , emit: contigs
-    path "versions.yml"  , emit: versions
+    tuple val(meta), path("contigs.bed") , emit: contigs
+    path "versions.yml"                  , emit: versions
 
     script: 
     def args       = task.ext.args ?: ''
     def prefix     = task.ext.prefix ?: "${meta.id}"
 
-    if (params.runcontigs == 'ALT_HLA')
+    if (contig_file)
     {
         """
-        if [ -n "\$(samtools view -H $tumor | grep -P "_alt\\tLN")" ]; then
-            samtools view -H $tumor | grep -P "_alt\\tLN" | sed -e 's/@SQ\\tSN://' -e 's/\\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 > ALT_contigs.bed
-        else
-            touch ALT_contigs.bed
-        fi
-        if [-n "\$(samtools view -H $tumor | grep -P "SN:HLA")" ]; then
-            samtools view -H $tumor | grep -P "SN:HLA" | sed -e 's/@SQ\\tSN://' -e 's/\\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 > HLA_contigs.bed
-        else
-            touch HLA_contigs.bed
-        fi
-        cat ALT_contigs.bed HLA_contigs.bed > contigs.bed
+        cp $contig_file contigs.bed
 
         cat <<-END_VERSIONS > versions.yml
         "${task.process}":
             samtools: \$(echo \$(samtools 2>&1) | sed -e 's/.*Version: //; s/ Usage.*//')
         END_VERSIONS
-
         """
     }
-    else {
-        """
-        touch contigs.bed
-        if [-n "\$(samtools view -H $tumor | grep -P "SN:") "]; then
-            samtools view -H $tumor | grep -P "SN:"  | sed -e 's/@SQ\\tSN://' -e 's/\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 | tail -n +25 > contigs.bed
-        else
+    else{
+        if (params.runcontigs == 'ALT_HLA')
+        {
+            """
+            if [ -n "\$(samtools view -H $tumor | grep -P "_alt\\tLN")" ]; then
+                samtools view -H $tumor | grep -P "_alt\\tLN" | sed -e 's/@SQ\\tSN://' -e 's/\\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 > ALT_contigs.bed
+            else
+                touch ALT_contigs.bed
+            fi
+            if [-n "\$(samtools view -H $tumor | grep -P "SN:HLA")" ]; then
+                samtools view -H $tumor | grep -P "SN:HLA" | sed -e 's/@SQ\\tSN://' -e 's/\\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 > HLA_contigs.bed
+            else
+                touch HLA_contigs.bed
+            fi
+            cat ALT_contigs.bed HLA_contigs.bed > contigs.bed
+
+            cat <<-END_VERSIONS > versions.yml
+            "${task.process}":
+                samtools: \$(echo \$(samtools 2>&1) | sed -e 's/.*Version: //; s/ Usage.*//')
+            END_VERSIONS
+
+            """
+        }
+        else {
+            """
             touch contigs.bed
-        fi
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            samtools: \$(echo \$(samtools 2>&1) | sed -e 's/.*Version: //; s/ Usage.*//')
-        END_VERSIONS
-        """
+            if [ -n "\$(samtools view -H $tumor | grep -P "SN:") "]; then
+                samtools view -H $tumor | grep -P "SN:"  | sed -e 's/@SQ\\tSN://' -e 's/\tLN:/\\t0\\t/' -e 's/\\tAH.*//' | sort -V -k1,1 | cut -f 1 | tail -n +25 > contigs.bed
+            else
+                touch contigs.bed
+            fi
+            cat <<-END_VERSIONS > versions.yml
+            "${task.process}":
+                samtools: \$(echo \$(samtools 2>&1) | sed -e 's/.*Version: //; s/ Usage.*//')
+            END_VERSIONS
+            """
+        }
+
     }
-}
+}
diff --git a/subworkflows/local/mpileup_snv_call.nf b/subworkflows/local/mpileup_snv_call.nf
@@ -15,28 +15,29 @@ workflow MPILEUP_SNV_CALL {
     sample_ch     // channel: [val(meta), tumor,tumor_bai, control, control_bai, tumorname, controlname]
     ref           // channel: [path(fasta), path(fai)]
     intervals     // channel: [[chr, region], [chr, region], ...]
-    contigs       // channel: [path(contigs.bed)]
+    contigs       // channel: [val(meta), path(contigs.bed)]
 
     main:
     versions = Channel.empty()
 
     // Combine intervals with samples to create 'interval x sample' number of parallel run and with contigs if exist
 
-    contigs.map { it -> ["contigs", it] }
+    contigs.map { it -> [it[0], "contigs", it[1],] }
         .set{contig_ch}
 
     intervals.take(24)
-            .map{it -> [it[0],[]]}
+            .map{it -> [[],it[0],[]]}
             .set{ch_intervals}
 
     ch_intervals.mix(contig_ch)
             .set{intervals_ch}
 
     sample_ch
             .combine(intervals_ch)
+            .filter{ it[0] != it[7] }
+            .map{ it -> [it[0], it[1], it[2], it[3], it[4], it[5], it[6], it[8], it[9] ] }
             .set { combined_inputs }
 
-    combined_inputs.view()        
     //
     // MODULE:BCFTOOLS_MPILEUP 
     //

diff --git a/workflows/snvcalling.nf b/workflows/snvcalling.nf
@@ -205,18 +205,18 @@ workflow SNVCALLING {
     }
     interval_ch  = chrlength.splitCsv(sep: '\t', by:1)
 
-    if ((params.runcontigs != "NONE") && (!params.contig_file)) {
+    if (params.runcontigs != "NONE") {
         //
         // MODULE: Prepare contigs file if not provided
         //
         GET_CONTIGS(
-            sample_ch
+            sample_ch,
+            contigs
             )
         ch_versions = ch_versions.mix(GET_CONTIGS.out.versions)
-        GET_CONTIGS.out.contigs.filter{contig -> WorkflowCommons.getNumLinesInFile(contig) > 0}
+        GET_CONTIGS.out.contigs.filter{meta, contig -> WorkflowCommons.getNumLinesInFile(contig) > 0}
                 .set{contigs}
     }
-
     //
     // MODULE: Extract sample name from BAM
     //