fixing bugs in bndeval outputs and ensemble analysis

CHANGELOG.md

@@ -29,6 +29,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - GT field is necessary for rtgtools vcfeval, strelka reports no GT field for somatic analysis. Fixing the GT field setup for strelka (bcftools plugingtset doesnt work as GT is not available) [#279](https://github.com/nf-core/variantbenchmarking/pull/279)
 - Using nf-metro to create better metromap [#281](https://github.com/nf-core/variantbenchmarking/pull/281)
 - Remove unneccesery reheadering of the vcf outputs of rtgtools and truvari [#286](https://github.com/nf-core/variantbenchmarking/pull/286)
+- Fixing bugs: bnd vcfeval output files producing tags with [] and reformatting header before merging in ensemble analysis [#297](https://github.com/nf-core/variantbenchmarking/pull/297)
 
 ### `Dependencies`
 

conf/modules.config

@@ -287,6 +287,14 @@ process {
         ]
     }
 
+    withName: BCFTOOLS_UNIFY_HEADER {
+        ext.prefix = {"${meta.id}.unify"}
+        ext.args   = {"-x FORMAT/AD,FORMAT/PL --output-type z --write-index=tbi"}
+        publishDir = [
+            enabled: false
+        ]
+    }
+
     withName: REFORMAT_TRUTH {
         ext.prefix = { input[0].baseName + '.reformatted' }
         ext.suffix = "vcf"

subworkflows/local/concordance_analysis/main.nf

@@ -53,8 +53,9 @@ workflow CONCORDANCE_ANALYSIS {
         }
 
     ch_bed_input = bed_ch
-        .map { file -> tuple(["id": "intervals"], file) }
-        .ifEmpty([[:], []])
+        .map { file -> [ [id: "intervals"], file ] }
+        .collect()
+        .ifEmpty( [ [id: "intervals"], [] ] )
 
     // GATK4 concordance does not support structural variants now - GATK4 SVCONCORDANCE is in beta
     GATK4_CONCORDANCE(

subworkflows/local/ensemble_test_vcfs/main.nf

@@ -13,6 +13,7 @@ include { GAWK as REFORMAT_TRUTH_SV           } from '../../../modules/nf-core/g
 include { GAWK as INJECT_MISSING_GT           } from '../../../modules/nf-core/gawk'
 include { BCFTOOLS_SORT as BCFTOOLS_SORT_SV   } from '../../../modules/nf-core/bcftools/sort'
 include { TABIX_BGZIP as TABIX_BGZIP_UNZIP    } from '../../../modules/nf-core/tabix/bgzip'
+include { BCFTOOLS_ANNOTATE as BCFTOOLS_UNIFY_HEADER } from '../../../modules/nf-core/bcftools/annotate'
 
 workflow ENSEMLE_TEST_VCFS {
     take:
@@ -22,8 +23,14 @@ workflow ENSEMLE_TEST_VCFS {
 
     main:
 
+    // unify header for callers
+    BCFTOOLS_UNIFY_HEADER(
+        test_vcfs.map{meta, vcf, index -> [meta, vcf, index, [], []]},
+        [],[],[]
+    )
+
     // if the benchmarking method is rtgtools, missing GT field is already filled in VCF preperation step, so no need to inject missing GT field
-    test_vcfs.branch { meta, vcf, index ->
+    BCFTOOLS_UNIFY_HEADER.out.vcf.join(BCFTOOLS_UNIFY_HEADER.out.tbi).branch { meta, vcf, index ->
         def is_rtg = params.method?.contains("rtgtools")
         def is_strelka_manta = ['strelka', 'manta'].contains(meta.caller.toLowerCase())
         def is_somatic = params.analysis == "somatic"

workflows/variantbenchmarking.nf

@@ -288,21 +288,13 @@ workflow VARIANTBENCHMARKING {
             ch_reports       = ch_reports.mix(RTGTOOLS_BNDEVAL.out.summary
                                                     .map { _meta, file -> tuple([vartype: params.variant_type] + [benchmark_tool: "rtgtools"], file) }
                                                     .groupTuple())
-            evals_ch         = evals_ch.mix(RTGTOOLS_BNDEVAL.out.fn_vcf,
-                                            RTGTOOLS_BNDEVAL.out.fp_vcf,
-                                            RTGTOOLS_BNDEVAL.out.baseline_vcf,
-                                            RTGTOOLS_BNDEVAL.out.tp_vcf)
-                                        .map { meta, file ->
-                                            def mapping = [
-                                                'fn': 'FN',
-                                                'fp': 'FP',
-                                                'tp-baseline': 'TP_base',
-                                                'tp': 'TP_comp'
-                                            ]
-                                            def tag = file.getName().tokenize('.').find { token -> token in ['fn', 'fp', 'tp-baseline', 'tp'] }
-                                            def transformedTag = mapping[tag] ?: tag
-                                            tuple( [ meta + [vartype: params.variant_type, id: "rtgtools", tag: transformedTag]], file)
-                                        }
+
+            evals_ch = evals_ch.mix(
+                RTGTOOLS_BNDEVAL.out.fn_vcf.map { _meta, file -> tuple([vartype: params.variant_type] + [tag: "FN"] + [id: "rtgtools"], file) },
+                RTGTOOLS_BNDEVAL.out.fp_vcf.map { _meta, file -> tuple([vartype: params.variant_type] + [tag: "FP"] + [id: "rtgtools"], file) },
+                RTGTOOLS_BNDEVAL.out.baseline_vcf.map { _meta, file -> tuple([vartype: params.variant_type] + [tag: "TP_base"] + [id: "rtgtools"], file) },
+                RTGTOOLS_BNDEVAL.out.tp_vcf.map { _meta, file -> tuple([vartype: params.variant_type] + [tag: "TP_comp"] + [id: "rtgtools"], file) }
+            )
         }
     }