Merge pull request #864 from maxulysse/dev_ultima

add realigned bam export possibilities to gatk4_haplotypecaller
nf-core · Nov 24, 2022 · 36bb01d · 36bb01d
2 parents 6c262b0 + 16ac8f6
commit 36bb01d
Show file tree

Hide file tree

Showing 7 changed files with 59 additions and 7 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -9,6 +9,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- [#864](https://github.com/nf-core/sarek/pull/864) - Added possibilities to export assembled haplotypes and locally realigned reads
+
 ### Changed
 
 - [#859](https://github.com/nf-core/sarek/pull/859) - Back to dev

diff --git a/conf/modules/haplotypecaller.config b/conf/modules/haplotypecaller.config
@@ -118,4 +118,16 @@ process {
             pattern: "*{vcf.gz,vcf.gz.tbi}"
         ]
     }
+
+    if (!(params.tools && params.tools.split(',').contains('haplotypecaller'))) {
+        withName: 'NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_HAPLOTYPECALLER:BAM_MERGE_INDEX_SAMTOOLS:(MERGE_BAM|INDEX_MERGE_BAM)' {
+            ext.prefix       = { "${meta.id}.realigned" }
+            publishDir       = [
+                enabled: true,
+                mode: params.publish_dir_mode,
+                path: { "${params.outdir}/variant_calling/haplotypecaller/${meta.id}/" },
+                pattern: "*{bam,bai}"
+            ]
+        }
+    }
 }
diff --git a/modules.json b/modules.json
@@ -12,7 +12,8 @@
                     },
                     "bcftools/mpileup": {
                         "branch": "master",
-                        "git_sha": "6301e29d77e7ec7ce98b55b8a361b316a9a91bfe"
+                        "git_sha": "6301e29d77e7ec7ce98b55b8a361b316a9a91bfe",
+                        "installed_by": ["modules"]
                     },
                     "bcftools/sort": {
                         "branch": "master",
@@ -233,7 +234,7 @@
                     },
                     "gatk4/haplotypecaller": {
                         "branch": "master",
-                        "git_sha": "7578005a576f3ac1b0d4f10c037dde34cdedfb67",
+                        "git_sha": "9e9cd1dc1865ee28514c6435e02f1618c091ae20",
                         "installed_by": ["modules"]
                     },
                     "gatk4/intervallisttobed": {

diff --git a/modules/nf-core/gatk4/haplotypecaller/main.nf b/modules/nf-core/gatk4/haplotypecaller/main.nf
diff --git a/modules/nf-core/gatk4/haplotypecaller/meta.yml b/modules/nf-core/gatk4/haplotypecaller/meta.yml
diff --git a/subworkflows/local/bam_variant_calling_haplotypecaller/main.nf b/subworkflows/local/bam_variant_calling_haplotypecaller/main.nf
@@ -1,7 +1,8 @@
-include { GATK4_MERGEVCFS as MERGE_HAPLOTYPECALLER } from '../../../modules/nf-core/gatk4/mergevcfs/main'
-include { GATK4_HAPLOTYPECALLER                    } from '../../../modules/nf-core/gatk4/haplotypecaller/main'
 include { BAM_JOINT_CALLING_GERMLINE_GATK          } from '../bam_joint_calling_germline_gatk/main'
+include { BAM_MERGE_INDEX_SAMTOOLS                 } from '../bam_merge_index_samtools/main'
 include { VCF_VARIANT_FILTERING_GATK               } from '../vcf_variant_filtering_gatk/main'
+include { GATK4_HAPLOTYPECALLER                    } from '../../../modules/nf-core/gatk4/haplotypecaller/main'
+include { GATK4_MERGEVCFS as MERGE_HAPLOTYPECALLER } from '../../../modules/nf-core/gatk4/mergevcfs/main'
 
 workflow BAM_VARIANT_CALLING_HAPLOTYPECALLER {
     take:
@@ -22,6 +23,7 @@ workflow BAM_VARIANT_CALLING_HAPLOTYPECALLER {
 
     ch_versions = Channel.empty()
     filtered_vcf = Channel.empty()
+    realigned_bam = Channel.empty()
 
     GATK4_HAPLOTYPECALLER(
         cram,
@@ -42,6 +44,11 @@ workflow BAM_VARIANT_CALLING_HAPLOTYPECALLER {
             no_intervals: it[0].num_intervals <= 1
         }.set{haplotypecaller_tbi_branch}
 
+    GATK4_HAPLOTYPECALLER.out.bam.branch{
+            intervals:    it[0].num_intervals > 1
+            no_intervals: it[0].num_intervals <= 1
+        }.set{haplotypecaller_bam_branch}
+
     if (params.joint_germline) {
         // merge vcf and tbis
         genotype_gvcf_to_call = Channel.empty().mix(GATK4_HAPLOTYPECALLER.out.vcf
@@ -100,6 +107,25 @@ workflow BAM_VARIANT_CALLING_HAPLOTYPECALLER {
             MERGE_HAPLOTYPECALLER.out.tbi,
             haplotypecaller_tbi_branch.no_intervals)
 
+        // BAM output
+        BAM_MERGE_INDEX_SAMTOOLS(
+            haplotypecaller_bam_branch.intervals
+            .map{ meta, bam ->
+
+                new_meta = [
+                                id:             meta.sample,
+                                num_intervals:  meta.num_intervals,
+                                patient:        meta.patient,
+                                sample:         meta.sample,
+                                sex:            meta.sex,
+                                status:         meta.status
+                            ]
+
+                    [groupKey(new_meta, new_meta.num_intervals), bam]
+                }.groupTuple().mix(haplotypecaller_bam_branch.no_intervals))
+
+        realigned_bam = BAM_MERGE_INDEX_SAMTOOLS.out.bam_bai
+
         VCF_VARIANT_FILTERING_GATK(haplotypecaller_vcf.join(haplotypecaller_tbi),
                     fasta,
                     fasta_fai,
@@ -117,4 +143,5 @@ workflow BAM_VARIANT_CALLING_HAPLOTYPECALLER {
     emit:
     versions = ch_versions
     filtered_vcf
+    realigned_bam
 }
diff --git a/tests/config/pytest_tags.yml b/tests/config/pytest_tags.yml
@@ -164,8 +164,11 @@ haplotypecaller:
   - modules/nf-core/gatk4/haplotypecaller/main.nf
   - modules/nf-core/gatk4/mergevcfs/main.nf
   - modules/nf-core/gatk4/variantrecalibrator/main.nf
+  - modules/nf-core/samtools/index/main.nf
+  - modules/nf-core/samtools/merge/main.nf
   - modules/nf-core/tabix/tabix/main.nf
   - subworkflows/local/bam_joint_calling_germline_gatk/main.nf
+  - subworkflows/local/bam_merge_index_samtools/main.nf
   - subworkflows/local/bam_variant_calling_haplotypecaller/main.nf
   - subworkflows/local/vcf_variant_filtering_gatk/main.nf