Merge branch 'dev' into software_version_fix

Author: jfy133

CHANGELOG.md

@@ -20,6 +20,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - [#750](https://github.com/nf-core/eager/issues/750) - Fixed piped commands requesting the same number of CPUs at each command step
 - [#757](https://github.com/nf-core/eager/issues/757) - Removed confusing 'Data Type' variable from MultiQC workflow summary (not consistent with TSV input)
 - [#759](https://github.com/nf-core/eager/pull/759) - Fixed malformed software scraping regex that resulted in N/A in MultiQC report
+- [#761](https://github.com/nf-core/eager/pull/759) - Fixed issues related to instability of samtools filtering related CI tests
 
 ### `Dependencies`
 

main.nf

@@ -1542,17 +1542,17 @@ process samtools_filter {
     // Using shell block rather than script because we are playing with awk
     script:
 
-    size = params.large_ref ? '-c' : ''
+    def size = params.large_ref ? '-c' : ''
 
     // Unmapped/MAPQ Filtering WITHOUT min-length filtering
     if ( "${params.bam_unmapped_type}" == "keep"  && params.bam_filter_minreadlength == 0 ) {
         """
-        samtools view -h -b ${bam} -@ ${task.cpus} -q ${params.bam_mapping_quality_threshold} > ${libraryid}.filtered.bam
+        samtools view -h ${bam} -@ ${task.cpus} -q ${params.bam_mapping_quality_threshold} -b > ${libraryid}.filtered.bam
         samtools index ${libraryid}.filtered.bam ${size}
         """
     } else if ( "${params.bam_unmapped_type}" == "discard" && params.bam_filter_minreadlength == 0 ){
         """
-        samtools view -h -b ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} > ${libraryid}.filtered.bam
+        samtools view -h ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} -b > ${libraryid}.filtered.bam
         samtools index ${libraryid}.filtered.bam ${size}
         """
     } else if ( "${params.bam_unmapped_type}" == "bam" && params.bam_filter_minreadlength == 0 ){
@@ -1563,14 +1563,9 @@ process samtools_filter {
         """
     } else if ( "${params.bam_unmapped_type}" == "fastq" && params.bam_filter_minreadlength == 0 ){
         """
-        echo "Samtools Filter Mapped"
-        samtools view -h ${bam} -@ ${task.cpus} -f4 -b -o ${libraryid}.unmapped.bam
-        echo "Samtools Filter Unmapped"
+        samtools view -h ${bam} -@ ${task.cpus} -f4 -b > ${libraryid}.unmapped.bam
         samtools view -h ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} -b > ${libraryid}.filtered.bam
-        echo "Samtools Indexing"
         samtools index ${libraryid}.filtered.bam ${size}
-
-        echo "Samtools BAM2FASTQ"
         ## FASTQ
         samtools fastq -tn ${libraryid}.unmapped.bam | pigz -p ${task.cpus - 1} > ${libraryid}.unmapped.fastq.gz
         rm ${libraryid}.unmapped.bam
@@ -1587,16 +1582,13 @@ process samtools_filter {
     // Unmapped/MAPQ Filtering WITH min-length filtering
     } else if ( "${params.bam_unmapped_type}" == "keep" && params.bam_filter_minreadlength != 0 ) {
         """
-        echo "Samtools quality filtering"
-        samtools view -h -b ${bam} -@ ${task.cpus} -q ${params.bam_mapping_quality_threshold} > tmp_mapped.bam
-        echo "Length filtering"
+        samtools view -h ${bam} -@ ${task.cpus} -q ${params.bam_mapping_quality_threshold} -b > tmp_mapped.bam
         filter_bam_fragment_length.py -a -l ${params.bam_filter_minreadlength} -o ${libraryid} tmp_mapped.bam
-        echo "Indexing"
         samtools index ${libraryid}.filtered.bam ${size}
         """
     } else if ( "${params.bam_unmapped_type}" == "discard" && params.bam_filter_minreadlength != 0 ){
         """
-        samtools view -h -b ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} > tmp_mapped.bam
+        samtools view -h ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} -b > tmp_mapped.bam
         filter_bam_fragment_length.py -a -l ${params.bam_filter_minreadlength} -o ${libraryid} tmp_mapped.bam
         samtools index ${libraryid}.filtered.bam ${size}
         """
@@ -1609,11 +1601,10 @@ process samtools_filter {
         """
     } else if ( "${params.bam_unmapped_type}" == "fastq" && params.bam_filter_minreadlength != 0 ){
         """
-        echo "Samtools Filter Mapped"
         samtools view -h ${bam} -@ ${task.cpus} -f4 -b > ${libraryid}.unmapped.bam
-        echo "Samtools Filter Unmapped"
-
         samtools view -h ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} -b > tmp_mapped.bam
+        filter_bam_fragment_length.py -a -l ${params.bam_filter_minreadlength} -o ${libraryid} tmp_mapped.bam
+        samtools index ${libraryid}.filtered.bam ${size}
 
         echo "Samtools Fragment Length Filtering"
         filter_bam_fragment_length.py -a -l ${params.bam_filter_minreadlength} -o ${libraryid} tmp_mapped.bam 
@@ -1628,15 +1619,11 @@ process samtools_filter {
         """
     } else if ( "${params.bam_unmapped_type}" == "both" && params.bam_filter_minreadlength != 0 ){
         """
-        echo "Samtools Filter Mapped"
         samtools view -h ${bam} -@ ${task.cpus} -f4 -b > ${libraryid}.unmapped.bam
-        echo "Samtools Filter Unmapped"
         samtools view -h ${bam} -@ ${task.cpus} -F4 -q ${params.bam_mapping_quality_threshold} -b > tmp_mapped.bam
-        echo "Samtools Fragment Length Filtering"
         filter_bam_fragment_length.py -a -l ${params.bam_filter_minreadlength} -o ${libraryid} tmp_mapped.bam
-        echo "Samtools Indexing"
         samtools index ${libraryid}.filtered.bam ${size}
-        echo "Samtools BAM2FASTQ"
+        
         ## FASTQ
         samtools fastq -tn ${libraryid}.unmapped.bam | pigz -p ${task.cpus} > ${libraryid}.unmapped.fastq.gz
         """