Merge branch 'CW-3188' into 'dev'

Fix joint phasing processing only a single chromosome

Closes CW-3188

See merge request epi2melabs/workflows/wf-human-variation!179

.gitlab-ci.yml

@@ -173,10 +173,10 @@ docker-run:
 
         - if: $MATRIX_NAME == "wf-human-phase_all"
           variables:
-              NF_BEFORE_SCRIPT: "wget -qO demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-str/demo_data.tar.gz && tar -xzvf demo_data.tar.gz"
+              NF_BEFORE_SCRIPT: "wget -q -O demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-variation/snp_demo.tar.gz && tar -xzvf demo_data.tar.gz"
               NF_PROCESS_FILES: "modules/local/wf-human-sv.nf modules/local/wf-human-snp.nf workflows/phasing.nf"
               NF_IGNORE_PROCESSES: "get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,hap,phase_contig,post_clair_phase_contig,bed_filter"
-              NF_WORKFLOW_OPTS: "--mod false --cnv false --str false --sv --snp --joint_phasing --bam demo_data/demo.bam --ref demo_data/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa.gz --sample_name demo --sex female --bam_min_coverage 0.00001 --annotation false"
+              NF_WORKFLOW_OPTS: "--mod false --cnv false --str false --sv --snp --joint_phasing --bam snp_demo/chr6_chr20.bam --bed snp_demo/chr6_chr20.bed --ref snp_demo/chr6_chr20.fasta --sample_name demo --sex female --bam_min_coverage 0.00001 --annotation false"
 
 
 aws-run:

CHANGELOG.md

@@ -5,6 +5,10 @@ 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).
 
 
+## [v1.9.2]
+### Fixed
+- `--joint_phasing` generating single-chromosome VCF files.
+
 ## [v1.9.1]
 ### Changed
 - ClinVar annotation of SVs has been temporarily removed due to not being correctly incorporated. SnpEff annotations are still produced as part of the final SV VCF.

nextflow.config

@@ -161,7 +161,7 @@ manifest {
     description     = 'Basecalling, SNV calling, SV calling, modified base calling, CNV calling, and STR genotyping of human samples.'
     mainScript      = 'main.nf'
     nextflowVersion = '>=23.04.2'
-    version         = '1.9.1'
+    version         = '1.9.2'
 }
 
 epi2melabs {

workflows/phasing.nf

@@ -8,10 +8,18 @@ process phase_all {
     // Phase VCF for a contig
     cpus 4
     input:
-        tuple path(snp_vcf, stageAs: "SNP/snp.vcf.gz"), path(snp_vcf_tbi, stageAs: "SNP/snp.vcf.gz.tbi")
-        tuple path(sv_vcf, stageAs: "SV/sv.vcf.gz"), path(sv_vcf_tbi, stageAs: "SV/sv.vcf.gz.tbi"), val(is_sv_vcf)
-        tuple path(ref), path(ref_idx), path(ref_cache), env(REF_PATH)
-        tuple val(contig), path(xam), path(xam_idx)
+        tuple val(contig),
+            path(xam),
+            path(xam_idx),
+            path(snp_vcf, stageAs: "SNP/snp.vcf.gz"),
+            path(snp_vcf_tbi, stageAs: "SNP/snp.vcf.gz.tbi"),
+            path(sv_vcf, stageAs: "SV/sv.vcf.gz"),
+            path(sv_vcf_tbi, stageAs: "SV/sv.vcf.gz.tbi"),
+            val(is_sv_vcf),
+            path(ref),
+            path(ref_idx),
+            path(ref_cache),
+            env(REF_PATH)
     output:
         path("phased_${contig}.vcf.gz"), emit: vcf
         path("phased_${contig}.vcf.gz.tbi"), emit: tbi
@@ -95,7 +103,10 @@ workflow phasing {
         }
 
         // Phase everything
-        phase_all(clair_vcf, struct_vcf, reference_ch, haplotagged_bams)
+        haplotagged_bams
+            .combine(clair_vcf)
+            .combine(struct_vcf)
+            .combine(reference_ch) | phase_all
 
         // Collect phased files
         vcfs = phase_all.out.vcf.collect()