Merge pull request GlobalPneumoSeq#139 from GlobalPneumoSeq/hotfix/li…

…mit-samtools-sort-max-thread

Cap maximum thread usage to samtools sort due to diminishing return and to avoid excessive memory request

bin/convert_sam_to_sorted_bam.sh

@@ -1,9 +1,13 @@
 # Convet SAM to sorted BAM file
 # Remove source SAM file if $LITE is true
 
-samtools view -@ "$(nproc)" -b "$SAM" > "$BAM"
+# Thread usage is capped as memory usage is per thread, and the speed gain level-off as thread count increases 
+AVAILABLE_THREAD=$(nproc)
+THREAD=$(( AVAILABLE_THREAD > MAX_THREAD ? MAX_THREAD : AVAILABLE_THREAD ))
 
-samtools sort -@ "$(nproc)" -o "$SORTED_BAM" "$BAM"
+samtools view -@ "$THREAD" -b "$SAM" > "$BAM"
+
+samtools sort -@ "$THREAD" -o "$SORTED_BAM" "$BAM"
 rm "$BAM"
 
 if [ "$LITE" = true ]; then

modules/mapping.nf

@@ -67,11 +67,13 @@ process SAM_TO_SORTED_BAM {
 
     script:
     sorted_bam="${sample_id}_mapped_sorted.bam"
+    max_thread="8"
     """
     SAM="$sam"
     BAM="mapped.bam"
     SORTED_BAM="$sorted_bam"
     LITE="$lite"
+    MAX_THREAD="$max_thread"
 
     source convert_sam_to_sorted_bam.sh
     source get_ref_coverage.sh