add skip for repeat annotation

CHANGELOG.md

@@ -7,13 +7,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### `Added`
 
+- Two new parameters `skip_repeat_annotation` and `skip_repeat_calling` to skip calling and annotation of repeat expansions [#574](https://github.com/nf-core/raredisease/pull/574)
 - A new parameter `skip_smncopynumbercaller` to skip smncopynumbercaller module[#574](https://github.com/nf-core/raredisease/pull/574)
 - A new parameter `skip_sv_calling` to skip sv calling workflow [#572](https://github.com/nf-core/raredisease/pull/572)
 - Two new parameters `skip_snv_calling` and `skip_repeat_analysis` to skip snv calling and repeat analysis respectively [#571](https://github.com/nf-core/raredisease/pull/571)
 - Two new parameters `mbuffer_mem` and `samtools_sort_threads` to control resources given to mbuffer and samtools sort in the bwameme module [#570](https://github.com/nf-core/raredisease/pull/570)
 
 ### `Changed`
 
+- Removed `skip_repeat_analysis` added in #571 [#574](https://github.com/nf-core/raredisease/pull/574)
 - Remove several skip parameters that had been included in the pipeline to avoid failed CI tests (see parameters table below) [#574](https://github.com/nf-core/raredisease/pull/574)
 - `readcount_intervals` parameter is now mandatory for running germlinecnvcaller. [#570](https://github.com/nf-core/raredisease/pull/570)
 - Turn off CNVnator, TIDDIT, SMNCopyNumberCaller, Gens, and Vcf2cytosure for targeted analysis [#573](https://github.com/nf-core/raredisease/pull/573)
@@ -29,14 +31,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 | --------------- | ------------------------ |
 |                 | mbuffer_mem              |
 |                 | samtools_sort_threads    |
-|                 | skip_repeat_analysis     |
+|                 | skip_repeat_calling      |
 |                 | skip_snv_calling         |
 |                 | skip_sv_calling          |
 | skip_eklipse    |                          |
 | skip_fastqc     |                          |
 | skip_haplocheck |                          |
 | skip_qualimap   |                          |
 |                 | skip_smncopynumbercaller |
+|                 | skip_repeat_annotation   |
 
 ## 2.1.0 - Obelix [2024-05-29]
 

conf/modules/annotate_repeat_expansions.config

@@ -0,0 +1,41 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Config file for defining DSL2 per module options and publishing paths
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    Available keys to override module options:
+        ext.args            = Additional arguments appended to command in module.
+        ext.args2           = Second set of arguments appended to command in module (multi-tool modules).
+        ext.args3           = Third set of arguments appended to command in module (multi-tool modules).
+        ext.prefix          = File name prefix for output files.
+        ext.when            = Conditional clause
+----------------------------------------------------------------------------------------
+*/
+
+//
+// Repeat expansion calling options
+//
+
+process {
+
+    withName: '.*ANNOTATE_REPEAT_EXPANSIONS:STRANGER' {
+        ext.prefix = { "${meta.id}_stranger" }
+    }
+
+    withName: '.*ANNOTATE_REPEAT_EXPANSIONS:COMPRESS_STRANGER' {
+        ext.prefix = { "${meta.id}_repeat_expansion_stranger" }
+        ext.args   = '--output-type z'
+        publishDir = [
+            path: { "${params.outdir}/repeat_expansions" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
+        ]
+    }
+
+    withName: '.*ANNOTATE_REPEAT_EXPANSIONS:INDEX_STRANGER' {
+        publishDir = [
+            path: { "${params.outdir}/repeat_expansions" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
+        ]
+    }
+}

conf/modules/call_repeat_expansions.config

@@ -58,25 +58,4 @@ process {
         ext.args = {"--notag"}
     }
 
-    withName: '.*CALL_REPEAT_EXPANSIONS:STRANGER' {
-        ext.prefix = { "${meta.id}_stranger" }
-    }
-
-    withName: '.*CALL_REPEAT_EXPANSIONS:COMPRESS_STRANGER' {
-        ext.prefix = { "${meta.id}_repeat_expansion_stranger" }
-        ext.args   = '--output-type z'
-        publishDir = [
-            path: { "${params.outdir}/repeat_expansions" },
-            mode: params.publish_dir_mode,
-            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
-        ]
-    }
-
-    withName: '.*CALL_REPEAT_EXPANSIONS:INDEX_STRANGER' {
-        publishDir = [
-            path: { "${params.outdir}/repeat_expansions" },
-            mode: params.publish_dir_mode,
-            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
-        ]
-    }
 }

nextflow.config

@@ -34,7 +34,8 @@ params {
     skip_me_calling            = false
     skip_me_annotation         = false
     skip_mt_annotation         = false
-    skip_repeat_analysis       = false
+    skip_repeat_annotation     = false
+    skip_repeat_calling        = false
     skip_smncopynumbercaller   = false
     skip_snv_annotation        = false
     skip_snv_calling           = false
@@ -316,6 +317,7 @@ includeConfig 'conf/modules/annotate_consequence_pli.config'
 includeConfig 'conf/modules/annotate_genome_snvs.config'
 includeConfig 'conf/modules/annotate_mt_snvs.config'
 includeConfig 'conf/modules/annotate_structural_variants.config'
+includeConfig 'conf/modules/annotate_repeat_expansions.config'
 includeConfig 'conf/modules/call_repeat_expansions.config'
 includeConfig 'conf/modules/call_snv.config'
 includeConfig 'conf/modules/call_structural_variants.config'

nextflow_schema.json

@@ -548,9 +548,14 @@
                     "description": "Specifies whether or not to subsample mt alignment.",
                     "fa_icon": "fas fa-toggle-on"
                 },
-                "skip_repeat_analysis": {
+                "skip_repeat_annotation": {
                     "type": "boolean",
-                    "description": "Specifies whether or not to skip calling and annotation of repeat expansions.",
+                    "description": "Specifies whether or not to skip annotation of repeat expansions.",
+                    "fa_icon": "fas fa-toggle-on"
+                },
+                "skip_repeat_calling": {
+                    "type": "boolean",
+                    "description": "Specifies whether or not to skip calling of repeat expansions.",
                     "fa_icon": "fas fa-toggle-on"
                 },
                 "skip_smncopynumbercaller": {

subworkflows/local/annotate_repeat_expansions.nf

@@ -0,0 +1,34 @@
+//
+// Annotate repeat expansions
+//
+
+include { BCFTOOLS_VIEW as COMPRESS_STRANGER           } from '../../modules/nf-core/bcftools/view/main'
+include { STRANGER                                     } from '../../modules/nf-core/stranger/main'
+include { TABIX_TABIX as INDEX_STRANGER                } from '../../modules/nf-core/tabix/tabix/main'
+
+workflow ANNOTATE_REPEAT_EXPANSIONS {
+    take:
+        ch_variant_catalog // channel: [mandatory] [ path(variant_catalog.json) ]
+        ch_vcf             // channel: [mandatory] [ val(meta), path(vcf) ]
+
+    main:
+        ch_versions = Channel.empty()
+
+        // Annotate, compress and index
+        STRANGER ( ch_vcf, ch_variant_catalog )
+        COMPRESS_STRANGER (
+            STRANGER.out.vcf.map{ meta, vcf -> [meta, vcf, [] ]},
+            [], [], []
+        )
+        INDEX_STRANGER ( COMPRESS_STRANGER.out.vcf )
+
+        ch_vcf_idx = COMPRESS_STRANGER.out.vcf.join(INDEX_STRANGER.out.tbi, failOnMismatch:true, failOnDuplicate:true)
+
+        ch_versions = ch_versions.mix(STRANGER.out.versions.first())
+        ch_versions = ch_versions.mix(COMPRESS_STRANGER.out.versions.first())
+        ch_versions = ch_versions.mix(INDEX_STRANGER.out.versions.first())
+
+emit:
+        vcf      = ch_vcf_idx   // channel: [ val(meta), path(vcf), path(tbi) ]
+        versions = ch_versions  // channel: [ path(versions.yml) ]
+}

subworkflows/local/call_repeat_expansions.nf

@@ -4,14 +4,11 @@
 
 include { BCFTOOLS_NORM as SPLIT_MULTIALLELICS_EXP     } from '../../modules/nf-core/bcftools/norm/main'
 include { BCFTOOLS_REHEADER as BCFTOOLS_REHEADER_EXP   } from '../../modules/nf-core/bcftools/reheader/main'
-include { BCFTOOLS_VIEW as COMPRESS_STRANGER           } from '../../modules/nf-core/bcftools/view/main'
 include { EXPANSIONHUNTER                              } from '../../modules/nf-core/expansionhunter/main'
 include { PICARD_RENAMESAMPLEINVCF as RENAMESAMPLE_EXP } from '../../modules/nf-core/picard/renamesampleinvcf/main'
-include { STRANGER                                     } from '../../modules/nf-core/stranger/main'
 include { SAMTOOLS_SORT                                } from '../../modules/nf-core/samtools/sort/main'
 include { SAMTOOLS_INDEX                               } from '../../modules/nf-core/samtools/index/main'
 include { SVDB_MERGE as SVDB_MERGE_REPEATS             } from '../../modules/nf-core/svdb/merge/main'
-include { TABIX_TABIX as INDEX_STRANGER                } from '../../modules/nf-core/tabix/tabix/main'
 include { TABIX_TABIX as TABIX_EXP_RENAME              } from '../../modules/nf-core/tabix/tabix/main'
 
 workflow CALL_REPEAT_EXPANSIONS {
@@ -63,29 +60,16 @@ workflow CALL_REPEAT_EXPANSIONS {
 
         SVDB_MERGE_REPEATS ( ch_svdb_merge_input, [] )
 
-        // Annotate, compress and index
-        STRANGER ( SVDB_MERGE_REPEATS.out.vcf,  ch_variant_catalog )
-        COMPRESS_STRANGER (
-            STRANGER.out.vcf.map{ meta, vcf -> [meta, vcf, [] ]},
-            [], [], []
-        )
-        INDEX_STRANGER ( COMPRESS_STRANGER.out.vcf )
-
-        ch_vcf_idx = COMPRESS_STRANGER.out.vcf.join(INDEX_STRANGER.out.tbi, failOnMismatch:true, failOnDuplicate:true)
-
         ch_versions = ch_versions.mix(EXPANSIONHUNTER.out.versions.first())
         ch_versions = ch_versions.mix(BCFTOOLS_REHEADER_EXP.out.versions.first())
         ch_versions = ch_versions.mix(RENAMESAMPLE_EXP.out.versions.first()    )
         ch_versions = ch_versions.mix(TABIX_EXP_RENAME.out.versions.first())
         ch_versions = ch_versions.mix(SPLIT_MULTIALLELICS_EXP.out.versions.first())
         ch_versions = ch_versions.mix(SVDB_MERGE_REPEATS.out.versions.first())
-        ch_versions = ch_versions.mix(STRANGER.out.versions.first())
-        ch_versions = ch_versions.mix(COMPRESS_STRANGER.out.versions.first())
-        ch_versions = ch_versions.mix(INDEX_STRANGER.out.versions.first())
         ch_versions = ch_versions.mix(SAMTOOLS_SORT.out.versions.first())
         ch_versions = ch_versions.mix(SAMTOOLS_INDEX.out.versions.first())
 
 emit:
-        vcf      = ch_vcf_idx  // channel: [ val(meta), path(vcf), path(tbi) ]
-        versions = ch_versions  // channel: [ path(versions.yml) ]
+        vcf      = SVDB_MERGE_REPEATS.out.vcf  // channel: [ val(meta), path(vcf) ]
+        versions = ch_versions                 // channel: [ path(versions.yml) ]
 }

workflows/raredisease.nf

@@ -31,7 +31,11 @@ if (params.run_rtgvcfeval) {
     mandatoryParams += ["rtg_truthvcfs"]
 }
 
-if (!params.skip_repeat_analysis) {
+if (!params.skip_repeat_calling) {
+    mandatoryParams += ["variant_catalog"]
+}
+
+if (!params.skip_repeat_annotation) {
     mandatoryParams += ["variant_catalog"]
 }
 
@@ -141,6 +145,7 @@ include { ANNOTATE_CSQ_PLI as ANN_CSQ_PLI_SV                 } from '../subworkf
 include { ANNOTATE_GENOME_SNVS                               } from '../subworkflows/local/annotate_genome_snvs'
 include { ANNOTATE_MOBILE_ELEMENTS                           } from '../subworkflows/local/annotate_mobile_elements'
 include { ANNOTATE_MT_SNVS                                   } from '../subworkflows/local/annotate_mt_snvs'
+include { ANNOTATE_REPEAT_EXPANSIONS                         } from '../subworkflows/local/annotate_repeat_expansions'
 include { ANNOTATE_STRUCTURAL_VARIANTS                       } from '../subworkflows/local/annotate_structural_variants'
 include { CALL_MOBILE_ELEMENTS                               } from '../subworkflows/local/call_mobile_elements'
 include { CALL_REPEAT_EXPANSIONS                             } from '../subworkflows/local/call_repeat_expansions'
@@ -448,7 +453,7 @@ workflow RAREDISEASE {
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 */
 
-    if (!params.skip_repeat_analysis && params.analysis_type.equals("wgs") ) {
+    if (!params.skip_repeat_calling && params.analysis_type.equals("wgs") ) {
         CALL_REPEAT_EXPANSIONS (
             ch_mapped.genome_bam_bai,
             ch_variant_catalog,
@@ -457,6 +462,14 @@ workflow RAREDISEASE {
             ch_genome_fai
         )
         ch_versions = ch_versions.mix(CALL_REPEAT_EXPANSIONS.out.versions)
+
+        if (!params.skip_repeat_annotation) {
+            ANNOTATE_REPEAT_EXPANSIONS (
+                ch_variant_catalog,
+                CALL_REPEAT_EXPANSIONS.out.vcf
+            )
+            ch_versions = ch_versions.mix(CALL_REPEAT_EXPANSIONS.out.versions)
+        }
     }