Compute index only once

nf-core · Jan 18, 2024 · 60052b6 · 60052b6
1 parent 533a0f3
commit 60052b6
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 49 deletions.
diff --git a/subworkflows/local/genome_quant.nf b/subworkflows/local/genome_quant.nf
@@ -8,32 +8,19 @@ include { BOWTIE_MAP_SEQ as BOWTIE_MAP_GENOME } from '../../modules/local/bowtie
 
 workflow GENOME_QUANT {
     take:
-    fasta
-    index
+    bowtie_index
+    fasta_formatted // fasta as generated by bowtie index step
     reads // channel: [ val(meta), [ reads ] ]
 
     main:
     ch_versions = Channel.empty()
 
-    if (!index){
-        INDEX_GENOME ( [ [:], fasta ] )
-        bowtie_index    = INDEX_GENOME.out.index
-        fasta_formatted = INDEX_GENOME.out.fasta
-        ch_versions     = ch_versions.mix(INDEX_GENOME.out.versions)
-    } else {
-        bowtie_index    = Channel.fromPath("${index}**ebwt", checkIfExists: true).ifEmpty { exit 1, "Bowtie1 index directory not found: ${index}" }
-        fasta_formatted = fasta
-    }
+    BOWTIE_MAP_GENOME ( reads, bowtie_index.collect() )
+    ch_versions = ch_versions.mix(BOWTIE_MAP_GENOME.out.versions)
 
-    if (bowtie_index){
-        BOWTIE_MAP_GENOME ( reads, bowtie_index.collect() )
-        ch_versions = ch_versions.mix(BOWTIE_MAP_GENOME.out.versions)
-        fasta_formatted
-            .map { file -> tuple(file.baseName, file) }
-            .set { sort_input }
-        BAM_SORT_STATS_SAMTOOLS ( BOWTIE_MAP_GENOME.out.bam,  sort_input )
-        ch_versions = ch_versions.mix(BAM_SORT_STATS_SAMTOOLS.out.versions)
-    }
+    ch_fasta_formatted_for_sort = fasta_formatted .map { file -> tuple(file.baseName, file) }
+    BAM_SORT_STATS_SAMTOOLS ( BOWTIE_MAP_GENOME.out.bam,  ch_fasta_formatted_for_sort )
+    ch_versions = ch_versions.mix(BAM_SORT_STATS_SAMTOOLS.out.versions)
 
     emit:
     fasta    = fasta_formatted

diff --git a/subworkflows/local/umi_dedup.nf b/subworkflows/local/umi_dedup.nf
@@ -12,7 +12,6 @@ include { CAT_CAT                             } from '../../modules/nf-core/cat/
 
 workflow DEDUPLICATE_UMIS {
     take:
-    fasta
     bt_index
     reads      // channel: [ val(meta), [ reads ] ]
     val_get_dedup_stats //boolean true/false
@@ -22,16 +21,6 @@ workflow DEDUPLICATE_UMIS {
     ch_versions = Channel.empty()
     ch_dedup_stats = Channel.empty()
 
-    if (!bt_index){
-        INDEX_GENOME ( [ [:], fasta ] )
-        bt_index      = INDEX_GENOME.out.index
-        fasta_formatted = INDEX_GENOME.out.fasta
-        ch_versions     = ch_versions.mix(INDEX_GENOME.out.versions)
-    } else {
-        fasta_formatted = fasta
-    }
-
-
     UMI_MAP_GENOME ( reads, bt_index.collect() )
     ch_versions = ch_versions.mix(UMI_MAP_GENOME.out.versions)
 

diff --git a/workflows/smrnaseq.nf b/workflows/smrnaseq.nf
@@ -150,29 +150,32 @@ workflow SMRNASEQ {
     )
     ch_versions = ch_versions.mix(FASTQ_FASTQC_UMITOOLS_FASTP.out.versions)
 
+    ch_fasta = file(params.fasta)
     reads_for_mirna = FASTQ_FASTQC_UMITOOLS_FASTP.out.reads
 
+    //Prepare bowtie index, unless specified
+    //This needs to be done here as the index is used by both UMI deduplication and GENOME_QUANT
+    if(params.bowtie_index) {
+        ch_bowtie_index  = Channel.fromPath("${index}**ebwt", checkIfExists: true).ifEmpty { error "Bowtie1 index directory not found: ${index}" }
+        ch_fasta_formatted = ch_fasta
+    } else {
+        INDEX_GENOME ( [ [:], ch_fasta ] )
+        ch_versions = ch_versions.mix(INDEX_GENOME.out.versions)
+        ch_bowtie_index = INDEX_GENOME.out.index
+        ch_fasta_formatted = INDEX_GENOME.out.fasta
+    }
+
     //
     // SUBWORKFLOW: Deduplicate UMIs by mapping them to the genome
     //
     if (params.with_umi){
-        if (params.fasta){
-            fasta_ch = file(params.fasta)
-
-            //This needs to be done here as GENOME_QUANT should not run prior to the deduplication of UMIs.
-            INDEX_GENOME ( [ [:], fasta_ch ] )
-
-            ch_versions = ch_versions.mix(INDEX_GENOME.out.versions)
-
-            DEDUPLICATE_UMIS (
-                fasta_ch,
-                INDEX_GENOME.out.index,
-                FASTQ_FASTQC_UMITOOLS_FASTP.out.reads,
-                params.umi_stats
-            )
-            reads_for_mirna = DEDUPLICATE_UMIS.out.reads
-            ch_versions = ch_versions.mix(DEDUPLICATE_UMIS.out.versions)
-        }
+        DEDUPLICATE_UMIS (
+            ch_bowtie_index,
+            FASTQ_FASTQC_UMITOOLS_FASTP.out.reads,
+            params.umi_stats
+        )
+        reads_for_mirna = DEDUPLICATE_UMIS.out.reads
+        ch_versions = ch_versions.mix(DEDUPLICATE_UMIS.out.versions)
     }
 
 
@@ -225,7 +228,7 @@ workflow SMRNASEQ {
     //
     genome_stats = Channel.empty()
     if (params.fasta){
-        GENOME_QUANT ( file(params.fasta), params.bowtie_index, MIRNA_QUANT.out.unmapped )
+        GENOME_QUANT ( ch_bowtie_index, ch_fasta_formatted, MIRNA_QUANT.out.unmapped )
         genome_stats = GENOME_QUANT.out.stats
         ch_versions = ch_versions.mix(GENOME_QUANT.out.versions)