Generic UMI handling subworkflow (#7208)

* Add starting files

* bam_dedup_umi working

* undo mistaken file change

* Add swf test config

* Add snapshot

subworkflows/nf-core/bam_dedup_umi/main.nf

@@ -0,0 +1,131 @@
+//
+// BAM deduplication with UMI processing
+//
+
+include { BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE as BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_TRANSCRIPTOME } from '../../../subworkflows/nf-core/bam_dedup_stats_samtools_umicollapse'
+include { BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS as BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_TRANSCRIPTOME       } from '../../../subworkflows/nf-core/bam_dedup_stats_samtools_umitools'
+include { BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE as BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_GENOME        } from '../../../subworkflows/nf-core/bam_dedup_stats_samtools_umicollapse'
+include { BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS as BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_GENOME              } from '../../../subworkflows/nf-core/bam_dedup_stats_samtools_umitools'
+include { BAM_SORT_STATS_SAMTOOLS                            } from '../../../subworkflows/nf-core/bam_sort_stats_samtools'
+include { UMITOOLS_PREPAREFORRSEM                            } from '../../../modules/nf-core/umitools/prepareforrsem'
+include { SAMTOOLS_SORT                                      } from '../../../modules/nf-core/samtools/sort/main'
+
+workflow BAM_DEDUP_UMI {
+    take:
+    ch_genome_bam         // channel: [ val(meta), path(bam), path(bai) ]
+    ch_fasta              // channel: [ val(meta), path(fasta) ]
+    umi_dedup_tool        // string: 'umicollapse' or 'umitools'
+    umitools_dedup_stats  // boolean: whether to generate UMI-tools dedup stats
+    bam_csi_index         // boolean: whether to generate CSI index
+    ch_transcriptome_bam  // channel: [ val(meta), path(bam) ]
+    ch_transcript_fasta   // channel: [ val(meta), path(fasta) ]
+
+    main:
+    ch_versions = Channel.empty()
+
+    if (umi_dedup_tool != "umicollapse" && umi_dedup_tool != "umitools"){
+        error("Unknown umi_dedup_tool '${umi_dedup_tool}'")
+    }
+
+    // Genome BAM deduplication
+    if (umi_dedup_tool == "umicollapse") {
+        BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_GENOME (
+            ch_genome_bam
+        )
+        UMI_DEDUP_GENOME = BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_GENOME
+        ch_dedup_log = UMI_DEDUP_GENOME.out.dedup_stats
+
+    } else if (umi_dedup_tool == "umitools") {
+        BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_GENOME (
+            ch_genome_bam,
+            umitools_dedup_stats
+        )
+        UMI_DEDUP_GENOME = BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_GENOME
+        ch_dedup_log = UMI_DEDUP_GENOME.out.deduplog
+    }
+
+    // Co-ordinate sort, index and run stats on transcriptome BAM. This takes
+    // some preparation- we have to coordinate sort the BAM, run the
+    // deduplication, then restore name sorting and run a script from umitools
+    // to prepare for rsem or salmon
+
+    // 1. Coordinate sort
+
+    BAM_SORT_STATS_SAMTOOLS (
+        ch_transcriptome_bam,
+        ch_transcript_fasta
+    )
+    ch_sorted_transcriptome_bam = BAM_SORT_STATS_SAMTOOLS.out.bam
+        .join(BAM_SORT_STATS_SAMTOOLS.out.bai)
+
+    // 2. Transcriptome BAM deduplication
+    if (umi_dedup_tool == "umicollapse") {
+        BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_TRANSCRIPTOME (
+            ch_sorted_transcriptome_bam
+        )
+        UMI_DEDUP_TRANSCRIPTOME = BAM_DEDUP_STATS_SAMTOOLS_UMICOLLAPSE_TRANSCRIPTOME
+        ch_dedup_log = ch_dedup_log.mix(UMI_DEDUP_GENOME.out.dedup_stats)
+
+    } else if (umi_dedup_tool == "umitools") {
+        BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_TRANSCRIPTOME (
+            ch_sorted_transcriptome_bam,
+            umitools_dedup_stats
+        )
+        UMI_DEDUP_TRANSCRIPTOME = BAM_DEDUP_STATS_SAMTOOLS_UMITOOLS_TRANSCRIPTOME
+        ch_dedup_log = ch_dedup_log.mix(UMI_DEDUP_GENOME.out.deduplog)
+    }
+
+    // 3. Restore name sorting
+    SAMTOOLS_SORT (
+        UMI_DEDUP_TRANSCRIPTOME.out.bam,
+        ch_fasta
+    )
+
+    // 4. Run prepare_for_rsem.py on paired-end BAM files
+    // This fixes paired-end reads in name sorted BAM files
+    // See: https://github.com/nf-core/rnaseq/issues/828
+    ended_transcriptome_dedup_bam = SAMTOOLS_SORT.out.bam
+        .branch {
+            meta, bam ->
+                single_end: meta.single_end
+                    return [ meta, bam ]
+                paired_end: !meta.single_end
+                    return [ meta, bam ]
+        }
+
+    UMITOOLS_PREPAREFORRSEM (
+        ended_transcriptome_dedup_bam.paired_end
+            .map { meta, bam -> [ meta, bam, [] ] }
+    )
+
+    ch_dedup_transcriptome_bam = ended_transcriptome_dedup_bam.single_end
+        .mix(UMITOOLS_PREPAREFORRSEM.out.bam)
+
+    // Collect files useful for MultiQC into one helpful emission. Don't
+    // automatically add transcriptome stats- difficult to separate in multiqc
+    // without a bit more work
+
+    ch_multiqc_files = ch_dedup_log
+        .mix(UMI_DEDUP_GENOME.out.stats)
+        .mix(UMI_DEDUP_GENOME.out.flagstat)
+        .mix(UMI_DEDUP_GENOME.out.idxstats)
+        .transpose()
+        .map{it[1]}
+
+    // Record versions
+
+    ch_versions = UMI_DEDUP_GENOME.out.versions
+        .mix(BAM_SORT_STATS_SAMTOOLS.out.versions)
+        .mix(UMITOOLS_PREPAREFORRSEM.out.versions)
+
+    emit:
+    bam                = UMI_DEDUP_GENOME.out.bam                                                // channel: [ val(meta), path(bam) ]
+    bai                = bam_csi_index ? UMI_DEDUP_GENOME.out.csi : UMI_DEDUP_GENOME.out.bai     // channel: [ val(meta), path(bai) ]
+    dedup_log          = ch_dedup_log                                                            // channel: [ val(meta), path(log) ]
+    stats              = UMI_DEDUP_GENOME.out.stats.mix(UMI_DEDUP_TRANSCRIPTOME.out.stats)       // channel: [ val(meta), path(stats)]
+    flagstat           = UMI_DEDUP_GENOME.out.flagstat.mix(UMI_DEDUP_TRANSCRIPTOME.out.flagstat) // channel: [ val(meta), path(flagstat)]
+    idxstats           = UMI_DEDUP_GENOME.out.idxstats.mix(UMI_DEDUP_TRANSCRIPTOME.out.idxstats) // channel: [ val(meta), path(idxstats)]
+    multiqc_files      = ch_multiqc_files                                                        // channel: file
+    transcriptome_bam  = ch_dedup_transcriptome_bam                                              // channel: [ val(meta), path(bam) ]
+    versions            = ch_versions                                                            // channel: [ path(versions.yml) ]
+}

subworkflows/nf-core/bam_dedup_umi/meta.yml

@@ -0,0 +1,181 @@
+name: "bam_dedup_umi"
+description: BAM deduplication with UMI processing for both genome and transcriptome alignments
+keywords:
+  - deduplication
+  - UMI
+  - BAM
+  - genome
+  - transcriptome
+  - umicollapse
+  - umitools
+
+components:
+  - umitools/prepareforrsem
+  - samtools/sort
+  - bam_dedup_stats_samtools_umicollapse
+  - bam_dedup_stats_samtools_umitools
+  - bam_sort_stats_samtools
+
+input:
+  - ch_genome_bam:
+      description: Channel with genome BAM files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - bam:
+            type: file
+            description: BAM file
+            pattern: "*.bam"
+        - bai:
+            type: file
+            description: BAM index file
+            pattern: "*.bai"
+  - ch_fasta:
+      description: Channel with genome FASTA file
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - fasta:
+            type: file
+            description: Genome FASTA file
+            pattern: "*.{fa,fasta}"
+  - umi_dedup_tool:
+      description: UMI deduplication tool to use
+      structure:
+        - value:
+            type: string
+            description: Either 'umicollapse' or 'umitools'
+  - umitools_dedup_stats:
+      description: Whether to generate UMI-tools deduplication stats
+      structure:
+        - value:
+            type: boolean
+            description: True or False
+  - bam_csi_index:
+      description: Whether to generate CSI index
+      structure:
+        - value:
+            type: boolean
+            description: True or False
+  - ch_transcriptome_bam:
+      description: Channel with transcriptome BAM files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - bam:
+            type: file
+            description: BAM file
+            pattern: "*.bam"
+  - ch_transcript_fasta:
+      description: Channel with transcript FASTA file
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - fasta:
+            type: file
+            description: Transcript FASTA file
+            pattern: "*.{fa,fasta}"
+
+output:
+  - bam:
+      description: Channel containing deduplicated genome BAM files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - bam:
+            type: file
+            description: Deduplicated BAM file
+            pattern: "*.bam"
+  - bai:
+      description: Channel containing indexed BAM (BAI) files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - bai:
+            type: file
+            description: BAM index file
+            pattern: "*.bai"
+  - csi:
+      description: Channel containing CSI files (if bam_csi_index is true)
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - csi:
+            type: file
+            description: CSI index file
+            pattern: "*.csi"
+  - dedup_log:
+      description: Channel containing deduplication log files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - log:
+            type: file
+            description: Deduplication log file
+            pattern: "*.log"
+  - stats:
+      description: Channel containing BAM statistics files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - stats:
+            type: file
+            description: BAM statistics file
+            pattern: "*.stats"
+  - flagstat:
+      description: Channel containing flagstat files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - flagstat:
+            type: file
+            description: Flagstat file
+            pattern: "*.flagstat"
+  - idxstats:
+      description: Channel containing idxstats files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - idxstats:
+            type: file
+            description: Idxstats file
+            pattern: "*.idxstats"
+  - multiqc_files:
+      description: Channel containing files for MultiQC
+      structure:
+        - file:
+            type: file
+            description: File for MultiQC
+  - transcriptome_bam:
+      description: Channel containing deduplicated transcriptome BAM files
+      structure:
+        - meta:
+            type: map
+            description: Metadata map
+        - bam:
+            type: file
+            description: Deduplicated transcriptome BAM file
+            pattern: "*.bam"
+  - versions:
+      description: Channel containing software versions file
+      structure:
+        - versions:
+            type: file
+            description: File containing versions of the software used
+            pattern: "versions.yml"
+
+authors:
+  - "@pinin4fjords"
+maintainers:
+  - "@pinin4fjords"