Merge branch 'CW-2493' into 'dev'

make output more practical [CW-2493]

See merge request epi2melabs/workflows/wf-amplicon!15

.gitlab-ci.yml

@@ -6,7 +6,7 @@ include:
 variables:
   NF_WORKFLOW_OPTS:
     "--fastq test_data/fastq --reference test_data/reference.fasta \
-    --threads 2"
+    --threads 2 --combine_results"
   CI_FLAVOUR: "new"
 
 docker-run:
@@ -30,7 +30,7 @@ docker-run:
       variables:
         NF_WORKFLOW_OPTS:
           "--fastq test_data/fastq --reference test_data/reference.fasta \
-          --threads 2 --reads_downsampling_size 2000"
+          --threads 2 --reads_downsampling_size 2000 --combine_results"
         NF_PROCESS_FILES: >
           main.nf
           modules/local/pipeline.nf
@@ -44,6 +44,7 @@ docker-run:
           --fastq test_data/fastq
           --reference test_data/reference.fasta
           --min_read_qual 20
+          --combine_results
         NF_PROCESS_FILES: >
           main.nf
           modules/local/pipeline.nf

CHANGELOG.md

@@ -4,6 +4,13 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [v0.3.0]
+### Changed
+- VCF files now use the sample alias as sample name instead of `SAMPLE`.
+- The reference FASTA file with sanitized sequence headers (with `:`, `*`, and whitespace replaced with `_`) which is used by the workflow internally due to some tools not tolerating these symbols in the sequence IDs is now also published alongside the other results.
+
+### Added
+- Parameter `--combine_results` to also output merged BAM and VCF files.
 
 ## [v0.2.3]
 ### Fixed

main.nf

@@ -221,13 +221,21 @@ workflow pipeline {
         | mix(
             software_versions | map { [it, null] },
             workflow_params | map { [it, null] },
+            variantCallingPipeline.out.sanitized_ref | map { [it, null] },
             variantCallingPipeline.out.variants
             | map { meta, vcf -> [vcf, "$meta.alias/variants"] },
             variantCallingPipeline.out.mapped
             | map { meta, bam -> [bam, "$meta.alias/alignments"] },
             variantCallingPipeline.out.consensus
             | map { meta, cons -> [cons, "$meta.alias/consensus"] },
         )
+        if (params.combine_results) {
+            ch_to_publish = ch_to_publish
+            | mix(
+                variantCallingPipeline.out.combined_vcfs | map { [it, null] },
+                variantCallingPipeline.out.combined_bams | map { [it, null] },
+            )
+        }
 
         makeReport(
             ch_results_for_report | collect,

modules/local/pipeline.nf

@@ -9,6 +9,7 @@ process alignReads {
     script:
     """
     minimap2 -t $task.cpus -ax map-ont reference.fasta reads.fastq.gz \
+        -R '@RG\\tID:$meta.alias\\tSM:$meta.alias' \
     | samtools sort -@ $task.cpus -o aligned.sorted.bam -
 
     samtools index aligned.sorted.bam
@@ -129,8 +130,9 @@ process medakaVariant {
     medaka tools annotate --dpsp medaka.vcf reference.fasta input.bam \
         medaka.annotated.unfiltered.vcf
 
-    # filter variants
-    bcftools filter medaka.annotated.unfiltered.vcf \
+    # use the sample alias as sample name in the VCF and filter variants
+    bcftools reheader medaka.annotated.unfiltered.vcf -s <(echo '$meta.alias') \
+    | bcftools filter \
         -e 'INFO/DP < $min_coverage' \
         -s LOW_DEPTH \
         -Oz -o medaka.annotated.vcf.gz
@@ -143,6 +145,35 @@ process medakaVariant {
     """
 }
 
+process mergeVCFs {
+    label "medaka"
+    cpus 1
+    input: path "VCFs/file*.vcf.gz"
+    output: path "combined.vcf.gz"
+    script:
+    """
+    (
+        cd VCFs
+        ls | xargs -n1 bcftools index
+    )
+    bcftools merge VCFs/file*.vcf.gz -Oz -o combined.vcf.gz
+    """
+}
+
+process mergeBAMs {
+    label "wfamplicon"
+    cpus 1
+    input:
+        path "BAMs/file*.bam"
+        path "indices/file*.bam.bai"
+    output: path "combined.bam"
+    script:
+    """
+    samtools merge BAMs/* indices/* -pXo combined.bam
+    """
+}
+
+
 // workflow module
 workflow pipeline {
     take:
@@ -196,10 +227,26 @@ workflow pipeline {
         mosdepth.out
         | groupTuple
         | concatMosdepthResultFiles
+
+        // combine per-sample BAM and VCF files if requested
+        combined_vcfs = null
+        combined_bams = null
+        if (params.combine_results) {
+            combined_vcfs = mergeVCFs(
+                medakaVariant.out.filtered.collect { meta, vcf -> vcf }
+            )
+            combined_bams = mergeBAMs(
+                alignReads.out.collect { meta, bam, bai -> bam },
+                alignReads.out.collect { meta, bam, bai -> bai }
+            )
+        }
     emit:
+        sanitized_ref = ref
         mapped = alignReads.out | map { meta, bam, bai -> [meta, bam] }
         mapping_stats = bamstats.out
         depth = concatMosdepthResultFiles.out
         variants = medakaVariant.out.filtered
         consensus = medakaVariant.out.consensus
+        combined_vcfs
+        combined_bams
 }

nextflow.config

@@ -18,6 +18,7 @@ params {
     sample = null
     sample_sheet = null
     analyse_unclassified = false
+    combine_results = false
 
     // filtering + downsampling args
     min_read_length = 300
@@ -65,7 +66,7 @@ manifest {
     description     = 'Amplicon workflow'
     mainScript      = 'main.nf'
     nextflowVersion = '>=23.04.2'
-    version         = 'v0.2.3'
+    version         = 'v0.3.0'
 }
 
 epi2melabs {

nextflow_schema.json

@@ -140,6 +140,12 @@
                     "default": "output",
                     "format": "directory-path",
                     "description": "Directory for output of all workflow results."
+                },
+                "combine_results": {
+                    "type": "boolean",
+                    "default": false,
+                    "description": "Whether to merge per-sample results into a single BAM / VCF file.",
+                    "help_text": "Per default, results are grouped per sample. With this option, an additional BAM and VCF file are produced which contain the alignments / variants for all samples and amplicons."
                 }
             }
         },