Features/lpa indepedent mode

conf/test.config

@@ -12,7 +12,6 @@ params {
     project                        = "exome-validation"
     input                          = "test-data/*.bam"
     gold                           = "reference-data/gold/gold.txt"
-    region                         = "reference-data/peterReadExtract.hg38.camo.LPA.realign.sorted.bed"
     reference                      = "reference-data/kiv2.fasta"
     contig                         = "KIV2_6"
     build                          = "hg38"

conf/test_other_regions.config

@@ -0,0 +1,18 @@
+/*
+========================================================================================
+    Nextflow config file for running minimal tests
+========================================================================================
+    Defines input files and everything required to run a fast and simple pipeline test.
+    Use as follows:
+        nextflow run genepi/exome-cnv-nf -r <tag> -profile test,<docker/singularity>
+----------------------------------------------------------------------------------------
+*/
+
+params {
+    project                        = "vntr-other-region"
+    input                          = "test-data/*.bam"
+    region                         = "bin/kiv2_hg38.bed"
+    reference                      = "reference-data/kiv2.fasta"
+    contig                         = "KIV2_6"
+    build                          = "hg38"
+}

modules/local/call_variants_mutserve.nf

@@ -4,7 +4,7 @@ process CALL_VARIANTS_MUTSERVE {
   input:
     path bamFiles
     path ref_fasta
-    path contig
+    val(contig)
 
   output:
     path "${params.project}.txt", emit: variants_ch

modules/local/extract_reads.nf

@@ -6,8 +6,8 @@ process EXTRACT_READS {
 
   output:
 	  path "*.extracted.bam", emit: extracted_bams_ch
-    
+
   """
-  samtools view -h -L ${regionFile} ${bamFile} | awk '\$5 < ${params.reads_quality_limit} || \$1 ~ "^@"' | samtools view -hb - | samtools sort -n -o ${bamFile.baseName}.extracted.bam -
+  samtools view -hb -L ${regionFile} ${bamFile} | samtools sort -n -o ${bamFile.baseName}.extracted.bam -
 	"""
 }

nextflow.config

@@ -20,6 +20,7 @@ params {
     gold                           = null
     region                         = null
     build                          = null
+    region                         = null
     reference                      = null
     signature_split_value          = 50
     mutserve_detection_limit       = 0.01

workflows/exome-cnv.nf

@@ -1,31 +1,7 @@
 nextflow.enable.dsl = 2
 
-params.project="exome-validation"
-params.input="$baseDir/test-data/*.bam"
-params.gold="$baseDir/reference-data/gold/gold.txt"
-params.region="$baseDir/reference-data/peterReadExtract.hg38.camo.LPA.realign.sorted.bed"
-params.reference="$baseDir/reference-data/kiv2.fasta"
-params.reference_fai="$baseDir/reference-data/kiv2.fasta.fai"
-params.contig="KIV2_6"
-params.threads = (Runtime.runtime.availableProcessors() - 1)
-params.output="output"
-
-bam_files_ch = Channel.fromPath(params.input)
-//region_file_ch = file(params.region)
-ref_fasta = file(params.reference)
-ref_fasta_fai = file(params.reference_fai)
-gold_standard = file(params.gold)
-contig = file(params.contig)
-
-if(params.build == "hg19") {
-lpa_region = file("$baseDir/bin/lpa_hg19.bed", checkIfExists: true)
-} else if (params.build == "hg38") {
-lpa_region = file("$baseDir/bin/lpa_hg38.bed", checkIfExists: true)
-}
-
 requiredParams = [
-    'project', 'input',
-    'gold', 'reference', 'contig', 'build'
+    'project', 'input','reference', 'contig'
 ]
 
 for (param in requiredParams) {
@@ -34,12 +10,17 @@ for (param in requiredParams) {
     }
 }
 
-if(params.outdir == null) {
-  outdir = "output/${params.project}"
-} else {
-  outdir = params.outdir
+if (params.region == null && params.build == null) {
+  exit 1, "Please specify build of your input data."
 }
 
+bam_files_ch = Channel.fromPath(params.input)
+ref_fasta = file(params.reference, checkIfExists: true)
+ref_fasta_fai = file(params.reference+".fai", checkIfExists: true)
+contig = params.contig
+
+outdir = "output/${params.project}"
+
 include { BUILD_BWA_INDEX       } from '../modules/local/build_bwa_index'
 include { DETECT_TYPE           } from '../modules/local/detect_type'  addParams(outdir: "$outdir")
 include { EXTRACT_READS         } from '../modules/local/extract_reads'  addParams(outdir: "$outdir")
@@ -52,10 +33,31 @@ include { CALCULATE_PERFORMANCE } from '../modules/local/calculate_performance'
 workflow EXOME_CNV {
 
     BUILD_BWA_INDEX ( ref_fasta )
+
+    // if no region is set, we apply our LPA-specific signature approach
+    if (params.region == null){
+      // load lpa region, concrecte bed file is then defined in DETECT_TYPE
+      if(params.build == "hg19") {
+        lpa_region = file("$baseDir/bin/lpa_hg19.bed", checkIfExists: true)
+      } else if (params.build == "hg38") {
+        lpa_region = file("$baseDir/bin/lpa_hg38.bed", checkIfExists: true)
+      }
     DETECT_TYPE ( bam_files_ch, lpa_region )
-    EXTRACT_READS ( DETECT_TYPE.out.bam_bed_ch )
+    bam_bed_tuple = DETECT_TYPE.out.bam_bed_ch
+    // for all other regions use the defined region file
+    } else {
+      region = file(params.region, checkIfExists: true)
+      bam_files_ch.map { it -> [it, region] }
+           .set { bam_bed_tuple }
+    }
+
+    EXTRACT_READS ( bam_bed_tuple )
     BAM_TO_FASTQ ( EXTRACT_READS.out.extracted_bams_ch )
     REALIGN_FASTQ ( BAM_TO_FASTQ.out.fastq_ch,ref_fasta,BUILD_BWA_INDEX.out.bwa_index_ch )
     CALL_VARIANTS_MUTSERVE ( REALIGN_FASTQ.out.realigned_ch.collect(),ref_fasta, contig )
-    CALCULATE_PERFORMANCE ( CALL_VARIANTS_MUTSERVE.out.variants_ch,gold_standard )
+
+    if(params.gold != null) {
+      gold_standard = file(params.gold, checkIfExists: true)
+      CALCULATE_PERFORMANCE ( CALL_VARIANTS_MUTSERVE.out.variants_ch,gold_standard )
+    }
 }