Merge pull request #30 from uclahs-cds/ts-regenotype

call-gSV Regenotyping
uclahs-cds · Jun 30, 2021 · 82e2cee · 82e2cee
2 parents 8358223 + 269cd0b
commit 82e2cee
Show file tree

Hide file tree

Showing 6 changed files with 151 additions and 45 deletions.
diff --git a/pipeline/call-gSV.nf b/pipeline/call-gSV.nf
@@ -51,7 +51,7 @@ Starting workflow...
 .stripIndent()
 
 include { run_validate } from './modules/validation'
-include { call_gSV_Delly; call_gCNV_Delly } from './modules/delly'
+include { call_gSV_Delly; call_gCNV_Delly; regenotype_gSV_Delly; regenotype_gCNV_Delly } from './modules/delly'
 include { call_gSV_Manta } from './modules/manta'
 include { convert_BCF2VCF_BCFtools as convert_gSV_BCF2VCF_BCFtools; convert_BCF2VCF_BCFtools as convert_gCNV_BCF2VCF_BCFtools } from './modules/bcftools'
 include { run_vcfstats_RTGTools as run_gSV_vcfstats_RTGTools; run_vcfstats_RTGTools as run_gCNV_vcfstats_RTGTools } from './modules/rtgtools'
@@ -65,7 +65,8 @@ input_bam_ch = Channel
                     row.patient,
                     row.sample,
                     row.input_bam,
-                    "${row.input_bam}.bai"
+                    "${row.input_bam}.bai",
+                    row.mode
                     )
         }
 
@@ -91,34 +92,46 @@ else {
     reference_fasta_index = "${params.reference_fasta}.fai"
     }
 
-// Create channel for validation
 validation_channel = Channel
     .fromPath(params.input_csv, checkIfExists:true)
     .splitCsv(header:true)
-    .map{ row -> [ 
-                row.input_bam,
-                params.reference_fasta
+    .map{ row -> [
+                'file-input',
+                row.input_bam
                 ]
         }
-    .flatten()
+
+Channel
+    .of(['file-fasta', params.reference_fasta])
+    .mix(validation_channel)
+    .set { validation_channel }
 
 workflow {
-    run_validate(validation_channel)
-    if (params.run_manta) {
-        call_gSV_Manta(input_bam_ch, params.reference_fasta, reference_fasta_index)
-        run_sha512sum_Manta(call_gSV_Manta.out.vcf_small_indel_sv_file.mix(call_gSV_Manta.out.vcf_diploid_sv_file, call_gSV_Manta.out.vcf_candidate_sv_file))
-        }
-    if (params.run_delly) {
-        call_gSV_Delly(input_bam_ch, params.reference_fasta, reference_fasta_index, params.exclusion_file)
-        call_gCNV_Delly(input_bam_ch, call_gSV_Delly.out.bcf_sv_file, params.reference_fasta, reference_fasta_index, params.mappability_map)
-        convert_gSV_BCF2VCF_BCFtools(call_gSV_Delly.out.bcf_sv_file, call_gSV_Delly.out.bam_sample_name, 'SV')
-        convert_gCNV_BCF2VCF_BCFtools(call_gCNV_Delly.out.bcf_cnv_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
-        if (params.run_qc) {
-            run_gSV_vcfstats_RTGTools(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gSV_Delly.out.bam_sample_name, 'SV')
-            run_gSV_vcf_validator_VCFtools(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gSV_Delly.out.bam_sample_name, 'SV')
-            run_gCNV_vcfstats_RTGTools(convert_gCNV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
-            run_gCNV_vcf_validator_VCFtools(convert_gCNV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
+    if (params.run_discovery) {
+        run_validate(validation_channel)
+        if (params.run_manta) {
+            call_gSV_Manta(input_bam_ch, params.reference_fasta, reference_fasta_index)
+            run_sha512sum_Manta(call_gSV_Manta.out.vcf_small_indel_sv_file.mix(call_gSV_Manta.out.vcf_diploid_sv_file, call_gSV_Manta.out.vcf_candidate_sv_file))
+            }
+        if (params.run_delly) {
+            call_gSV_Delly(input_bam_ch, params.reference_fasta, reference_fasta_index, params.exclusion_file)
+            call_gCNV_Delly(input_bam_ch, call_gSV_Delly.out.bcf_sv_file, params.reference_fasta, reference_fasta_index, params.mappability_map)
+            convert_gSV_BCF2VCF_BCFtools(call_gSV_Delly.out.bcf_sv_file, call_gSV_Delly.out.bam_sample_name, 'SV')
+            convert_gCNV_BCF2VCF_BCFtools(call_gCNV_Delly.out.bcf_cnv_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
+            if (params.run_qc) {
+                run_gSV_vcfstats_RTGTools(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gSV_Delly.out.bam_sample_name, 'SV')
+                run_gSV_vcf_validator_VCFtools(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gSV_Delly.out.bam_sample_name, 'SV')
+                run_gCNV_vcfstats_RTGTools(convert_gCNV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
+                run_gCNV_vcf_validator_VCFtools(convert_gCNV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bam_sample_name, 'CNV')
+                }
+            run_sha512sum_Delly(call_gSV_Delly.out.bcf_sv_file.mix(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bcf_cnv_file, convert_gCNV_BCF2VCF_BCFtools.out.vcf_file))
             }
-        run_sha512sum_Delly(call_gSV_Delly.out.bcf_sv_file.mix(convert_gSV_BCF2VCF_BCFtools.out.vcf_file, call_gCNV_Delly.out.bcf_cnv_file, convert_gCNV_BCF2VCF_BCFtools.out.vcf_file))
         }
+    // When 'run_regenotyping' is set to true, the mode specified in the input_csv will be used to determine which
+    // regenotyping process to run. For example, if the mode contains 'SV', regenotype_gSV_Delly will run, etc.
+    if (params.run_regenotyping) {
+        run_validate(validation_channel)
+        regenotype_gSV_Delly(input_bam_ch, params.reference_fasta, reference_fasta_index, params.exclusion_file, params.merged_sites)
+        regenotype_gCNV_Delly(input_bam_ch, params.reference_fasta, reference_fasta_index, params.exclusion_file, params.merged_sites)
+    }
 }
diff --git a/pipeline/config/nextflow.config b/pipeline/config/nextflow.config
@@ -16,6 +16,12 @@ params {
     blcds_registered_dataset = false
     sge_scheduler = false
 
+    // The mode of the call-gSV pipeline to run. Discovery will identify SVs using Delly/Manta. Regenotyping will regenotype SVs/CNVs using Delly. Please note that setting
+    // `run_regenotyping` to true will supersede `run_delly` below.
+    run_discovery = true
+    run_regenotyping = false
+    merged_sites = '/path/to/sites.bcf'
+
     input_csv = 'path/to/input.csv'
     reference_fasta = '/path/to/genome.fa'
     // If an index is not explicitly specified for the reference FASTA, the pipeline will assume one exists in the same directory as the reference FASTA

diff --git a/pipeline/inputs/call-gSV-inputs.csv b/pipeline/inputs/call-gSV-inputs.csv
@@ -1,2 +1,2 @@
-patient,sample,input_bam
-patient-name,sample-name,/path/to/sample.bam
+patient,sample,input_bam,mode
+patient-name,sample-name,/path/to/sample.bam,mode
diff --git a/pipeline/modules/delly.nf b/pipeline/modules/delly.nf
@@ -25,17 +25,17 @@ process call_gSV_Delly {
         saveAs: { "call_gSV_Delly/${bam_sample_name}.log${file(it).getName()}" }
 
     input:
-    tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai)
-    path(reference_fasta)
-    path(reference_fasta_fai)
-    path(exclusion_file)
+        tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai), val(mode)
+        path(reference_fasta)
+        path(reference_fasta_fai)
+        path(exclusion_file)
 
 
     output:
-    path "DELLY-${params.delly_version}_SV_${params.dataset_id}_${bam_sample_name}.bcf", emit: bcf_sv_file
-    path "DELLY-${params.delly_version}_SV_${params.dataset_id}_${bam_sample_name}.bcf.csi"
-    path ".command.*"
-    val bam_sample_name, emit: bam_sample_name
+        path "DELLY-${params.delly_version}_SV_${params.dataset_id}_${bam_sample_name}.bcf", emit: bcf_sv_file
+        path "DELLY-${params.delly_version}_SV_${params.dataset_id}_${bam_sample_name}.bcf.csi"
+        path ".command.*"
+        val bam_sample_name, emit: bam_sample_name
 
     """
     set -euo pipefail
@@ -49,6 +49,49 @@ process call_gSV_Delly {
     """
 }
 
+process regenotype_gSV_Delly {
+    container docker_image_delly
+
+    publishDir params.output_dir,
+        enabled: params.save_intermediate_files,
+        pattern: "*.bcf*",
+        mode: "copy",
+        saveAs: { "Delly-${params.delly_version}/${file(it).getName()}" }
+
+    publishDir params.output_log_dir,
+        pattern: ".command.*",
+        mode: "copy",
+        saveAs: { "regenotype_gSV_Delly/${bam_sample_name}.log${file(it).getName()}" }
+
+    input:
+        tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai), val(mode)
+        path(reference_fasta)
+        path(reference_fasta_fai)
+        path(exclusion_file_or_mappability_map)
+        path(sites)
+
+    when:
+        mode == 'SV'
+
+    output:
+        path "DELLY-${params.delly_version}_RGSV_${params.dataset_id}_${bam_sample_name}.bcf", emit: regenotyped_sv_bcf
+        path "DELLY-${params.delly_version}_RGSV_${params.dataset_id}_${bam_sample_name}.bcf.csi", emit: regenotyped_sv_bcf_csi
+        path ".command.*"
+
+    script:
+    """
+    set -euo pipefail
+    delly \
+        call \
+        --vcffile $sites \
+        --exclude $exclusion_file_or_mappability_map \
+        --genome $reference_fasta \
+        --outfile "DELLY-${params.delly_version}_RGSV_${params.dataset_id}_${bam_sample_name}.bcf" \
+        --map-qual ${params.map_qual} \
+        "$input_bam"
+    """
+}
+
 process call_gCNV_Delly {
     container docker_image_delly
 
@@ -64,17 +107,17 @@ process call_gCNV_Delly {
         saveAs: { "call_gCNV_Delly/${bam_sample_name}.log${file(it).getName()}" }
 
     input:
-    tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai)
-    path(delly_sv_file)
-    path(reference_fasta)
-    path(reference_fasta_fai)
-    path(mappability_file)
+        tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai), val(mode)
+        path(delly_sv_file)
+        path(reference_fasta)
+        path(reference_fasta_fai)
+        path(mappability_file)
 
     output:
-    path "DELLY-${params.delly_version}_CNV_${params.dataset_id}_${bam_sample_name}.bcf", emit: bcf_cnv_file
-    path "DELLY-${params.delly_version}_CNV_${params.dataset_id}_${bam_sample_name}.bcf.csi"
-    path ".command.*"
-    val bam_sample_name, emit: bam_sample_name
+        path "DELLY-${params.delly_version}_CNV_${params.dataset_id}_${bam_sample_name}.bcf", emit: bcf_cnv_file
+        path "DELLY-${params.delly_version}_CNV_${params.dataset_id}_${bam_sample_name}.bcf.csi"
+        path ".command.*"
+        val bam_sample_name, emit: bam_sample_name
 
     """
     set -euo pipefail
@@ -87,3 +130,46 @@ process call_gCNV_Delly {
         $input_bam
     """
 }
+
+process regenotype_gCNV_Delly {
+    container docker_image_delly
+
+    publishDir params.output_dir,
+        enabled: params.save_intermediate_files,
+        pattern: "*.bcf*",
+        mode: "copy",
+        saveAs: { "Delly-${params.delly_version}/${file(it).getName()}" }
+
+    publishDir params.output_log_dir,
+        pattern: ".command.*",
+        mode: "copy",
+        saveAs: { "regenotype_gCNV_Delly/${bam_sample_name}.log${file(it).getName()}" }
+
+    input:
+        tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai), val(mode)
+        path(reference_fasta)
+        path(reference_fasta_fai)
+        path(exclusion_file_or_mappability_map)
+        path(sites)
+
+    output:
+        path "DELLY-${params.delly_version}_RGCNV_${params.dataset_id}_${bam_sample_name}.bcf", emit: regenotyped_cnv_bcf
+        path "DELLY-${params.delly_version}_RGCNV_${params.dataset_id}_${bam_sample_name}.bcf.csi", emit: regenotyped_cnv_bcf_csi
+        path ".command.*"
+
+    when:
+        mode == 'CNV'
+
+    script:
+    """
+    set -euo pipefail
+    delly \
+        cnv \
+        --segmentation \
+        --vcffile       $sites \
+        --exclude $exclusion_file_or_mappability_map \
+        --genome $reference_fasta \
+        --outfile "DELLY-${params.delly_version}_RGCNV_${params.dataset_id}_${bam_sample_name}.bcf" \
+        "$input_bam"
+    """
+}
diff --git a/pipeline/modules/manta.nf b/pipeline/modules/manta.nf
@@ -24,7 +24,7 @@ process call_gSV_Manta {
         saveAs: { "call_gSV_Manta/${bam_sample_name}.log${file(it).getName()}" }
 
     input:
-    tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai)
+    tuple val(patient), val(bam_sample_name), path(input_bam), path(input_bam_bai), val(mode)
     path(reference_fasta)
     path(reference_fasta_fai)
 

diff --git a/pipeline/modules/validation.nf b/pipeline/modules/validation.nf
@@ -19,13 +19,14 @@ process run_validate {
         saveAs: { "run_validate/${file_to_validate}.log${file(it).getName()}" }
 
     input:
-    path(file_to_validate)
+    tuple val(mode), path(file_to_validate)
 
     output:
+    path("${file_to_validate}.temp")
     path ".command.*"
 
     """
     set -euo pipefail
-    python -m validate -t file-input ${file_to_validate}
+    python -m validate -t ${mode} ${file_to_validate} > "${file_to_validate}.temp"
     """
 }