Fgumi initial commit

assets/schema_input.json

@@ -53,6 +53,24 @@
                 "description": "Run markdup in UMI-aware mode. This applies to Samtools only and requires the UMI to be in the read name.",
                 "default": false
             },
+            "fgumi_aware": {
+                "meta": ["fgumi_aware"],
+                "type": "boolean",
+                "description": "Enable UMI-aware consensus processing through the fgumi branch.",
+                "default": false
+            },
+            "fgumi_read_structures": {
+                "meta": ["fgumi_read_structures"],
+                "type": "string",
+                "description": "Read structures passed to fgumi extract for this sample.",
+                "default": null
+            },
+            "fgumi_extract_umis_from_read_names": {
+                "meta": ["fgumi_extract_umis_from_read_names"],
+                "type": "boolean",
+                "description": "Override fgumi extraction from read names for this sample.",
+                "default": null
+            },
             "skip_trimming": {
                 "meta": ["skip_trimming"],
                 "type": "boolean",

assets/schema_sampleinfo.json

@@ -93,6 +93,24 @@
                 "description": "Run markdup in UMI-aware mode. This applies to Samtools only and requires the UMI to be in the read name.",
                 "default": false
             },
+            "fgumi_aware": {
+                "meta": ["fgumi_aware"],
+                "type": "boolean",
+                "description": "Enable UMI-aware consensus processing through the fgumi branch.",
+                "default": false
+            },
+            "fgumi_read_structures": {
+                "meta": ["fgumi_read_structures"],
+                "type": "string",
+                "description": "Read structures passed to fgumi extract for this sample.",
+                "default": null
+            },
+            "fgumi_extract_umis_from_read_names": {
+                "meta": ["fgumi_extract_umis_from_read_names"],
+                "type": "boolean",
+                "description": "Override fgumi extraction from read names for this sample.",
+                "default": null
+            },
             "skip_trimming": {
                 "meta": ["skip_trimming"],
                 "type": "boolean",

conf/modules.config

@@ -231,6 +231,127 @@ process {
         }
     }
 
+    //// FGUMI extract (step 1)
+    withName: '.*FGUMI_EXTRACT' {
+        cpus       = 4
+        memory     = 16.GB
+        ext.prefix = { "${meta.id}.fgumi.unmapped" }
+        ext.args   = {
+            [
+                "--sample \"${meta.id}\"",
+                "--library \"${meta.library ?: meta.id}\"",
+                "--read-structures ${meta.fgumi_read_structures ?: '+T +T'}",
+                ((meta.fgumi_extract_umis_from_read_names != null ? meta.fgumi_extract_umis_from_read_names : true) ? "--extract-umis-from-read-names" : ""),
+                "--compression-level ${params.fgumi_compression_level}",
+            ].join(" ").trim()
+        }
+    }
+
+    //// FGUMI fastq | SNAP | zipper | template sort (step 3)
+    withName: '.*FGUMI_SNAP_ZIPPER_SORT' {
+        cpus       = 16
+        memory     = 64.GB
+        ext.prefix = { "${meta.id}.fgumi" }
+        ext.args   = {
+            [
+                "-b-",
+                "-sm 20",
+                params.fgumi_snap_ignore_mismatched_pairs ? "-I" : "",
+                "-hc-",
+                "-S id",
+                "-sa",
+                "-xf 2",
+                meta.readgroup ? "-R \"@RG\\t" + meta.readgroup.findResults { rg -> rg.value?.trim() ? "${rg.key}:${rg.value}" : null }.join("\\t") + "\"" : "",
+            ].join(" ").trim()
+        }
+        ext.args2  = {
+            [
+                "--threads ${task.cpus}",
+            ].join(" ").trim()
+        }
+        ext.args3  = {
+            [
+                "--threads ${task.cpus}",
+                "--max-memory ${params.fgumi_sort_max_memory}",
+                "--memory-per-thread=${params.fgumi_sort_memory_per_thread ? 'true' : 'false'}",
+            ].join(" ").trim()
+        }
+        ext.args4  = {
+            [
+                "--threads ${task.cpus}",
+                "--max-memory ${params.fgumi_sort_max_memory}",
+                "--memory-per-thread=${params.fgumi_sort_memory_per_thread ? 'true' : 'false'}",
+            ].join(" ").trim()
+        }
+        ext.args5  = {
+            [
+                "-@ ${task.cpus}",
+            ].join(" ").trim()
+        }
+    }
+
+    //// FGUMI group (step 4)
+    withName: '.*FGUMI_GROUP' {
+        cpus       = 8
+        memory     = 32.GB
+        ext.prefix = { "${meta.id}.fgumi.group" }
+        ext.args   = {
+            [
+                "--strategy ${meta.fgumi_group_strategy ?: 'adjacency'}",
+                "--edits ${meta.fgumi_group_edits != null ? meta.fgumi_group_edits : 1}",
+                "--compression-level ${meta.fgumi_compression_level != null ? meta.fgumi_compression_level : 1}",
+                "--grouping-metrics ${meta.id}.fgumi.group.grouping_metrics.txt",
+                "--family-size-histogram ${meta.id}.fgumi.group.family_size_histogram.txt",
+            ].join(" ").trim()
+        }
+    }
+
+    //// FGUMI simplex (step 5)
+    withName: '.*FGUMI_SIMPLEX' {
+        cpus       = 8
+        memory     = 32.GB
+        ext.prefix = { "${meta.id}.fgumi.simplex" }
+        ext.args   = {
+            [
+                "--min-reads ${params.fgumi_simplex_min_reads}",
+                "--threads ${task.cpus}",
+                "--queue-memory ${params.fgumi_queue_memory}",
+                params.fgumi_queue_memory_per_thread ? "--queue-memory-per-thread" : "",
+                "--compression-level ${params.fgumi_compression_level}",
+                "--stats ${meta.id}.fgumi.simplex.consensus_metrics.txt",
+            ].join(" ").trim()
+        }
+    }
+
+    //// FGUMI filter + coordinate sort/index (step 7)
+    withName: '.*FGUMI_FILTER' {
+        cpus       = 4
+        memory     = 16.GB
+        ext.prefix = { "${meta.id}.fgumi.filter" }
+        ext.args   = {
+            [
+                "--min-reads 1,1,1",
+                "--stats ${meta.id}.fgumi.filter.filtering_metrics.txt",
+            ].join(" ").trim()
+        }
+    }
+
+    //// FGUMI coordinate sort/index after filter (step 7)
+    withName: '.*FGUMI_SORT' {
+        cpus       = 8
+        memory     = 32.GB
+        ext.prefix = { "${meta.id}.fgumi.filter" }
+        ext.args   = {
+            [
+                "--order coordinate",
+                "--write-index",
+                "--threads ${task.cpus}",
+                "--max-memory ${params.fgumi_sort_max_memory}",
+                "--memory-per-thread=${params.fgumi_sort_memory_per_thread ? 'true' : 'false'}",
+            ].join(" ").trim()
+        }
+    }
+
     // coverage
     //// Mosdepth
     withName: '.*COVERAGE:MOSDEPTH' {

docs/usage.md

@@ -26,6 +26,7 @@ A `fastq` samplesheet file consisting of paired-end data may look something like
   aligner: bwamem
   markdup: bamsormadup
   umi_aware: false
+  fgumi_aware: false
   skip_trimming: false
   trim_front: 0
   trim_tail: 0
@@ -67,6 +68,10 @@ Following table shows the fields that are used by the `fastq` samplesheet:
 
 An [example samplesheet](../tests/inputs/test.yml) has been provided with the pipeline.
 
+> [!NOTE]
+> `umi_aware` and `fgumi_aware` are independent options.
+> Use `umi_aware` for samtools markdup UMI mode, and `fgumi_aware` to run the fgumi consensus branch.
+
 ### Flowcell samplesheet
 
 A `flowcell` samplesheet file consisting of one sequencing run may look something like the one below.
@@ -102,6 +107,7 @@ A `flowcell` sample info JSON/YML file consisting for one sequencing run may loo
   aligner: bwamem
   markdup: bamsormadup
   umi_aware: false
+  fgumi_aware: false
   skip_trimming: false
   trim_front: 0
   trim_tail: 0

main.nf

@@ -79,6 +79,9 @@ workflow {
     rna_junctions              = PREPROCESSING.out.rna_junctions
     align_reports              = PREPROCESSING.out.align_reports
     sormadup_metrics           = PREPROCESSING.out.sormadup_metrics
+    // Additional UMI consensus outputs.
+    family_size_histogram      = PREPROCESSING.out.family_size_histogram
+    umi_filtered_consensus_bam = PREPROCESSING.out.umi_filtered_consensus_bam
     mosdepth_global            = PREPROCESSING.out.mosdepth_global
     mosdepth_summary           = PREPROCESSING.out.mosdepth_summary
     mosdepth_regions           = PREPROCESSING.out.mosdepth_regions
@@ -176,6 +179,17 @@ output {
             metrics >> (meta.library ? "${meta.library}/${meta.samplename}/${meta.samplename}.duplicate_metrics.txt" : "${meta.samplename}/${meta.samplename}.duplicate_metrics.txt")
         }
     }
+    // UMI consensus artefacts are published per sample next to CRAM outputs.
+    family_size_histogram {
+        path { meta, histogram ->
+            histogram >> (meta.library ? "${meta.library}/${meta.samplename}/${histogram.name}" : "${meta.samplename}/${histogram.name}")
+        }
+    }
+    umi_filtered_consensus_bam {
+        path { meta, bam ->
+            bam >> (meta.library ? "${meta.library}/${meta.samplename}/${bam.name}" : "${meta.samplename}/${bam.name}")
+        }
+    }
     mosdepth_global {
         path { meta, _file ->
             return (meta.library ? "${meta.library}/${meta.samplename}/" : "${meta.samplename}/")

modules/local/fgumi/extract/main.nf

@@ -0,0 +1,40 @@
+process FGUMI_EXTRACT {
+    tag "$meta.id"
+    label 'process_medium'
+
+    container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
+        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/95/954170443a820787c9e02ef2135ebb8ec29c6b03633b0d61b5fafa98c59a1cce/data'
+        : 'community.wave.seqera.io/library/fgumi_r-base_r-ggplot2_r-scales:09c99070b82c1c28'}"
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("${prefix}.bam"), emit: bam
+    tuple val("${task.process}"), val('fgumi'), eval("fgumi --version | sed 's/^fgumi //;q'"), topic: versions, emit: versions_fgumi
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    // Derive per-thread queue memory from requested process resources.
+    def queue_memory_mb = (task.memory.mega / task.cpus * 0.75).intValue()
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.unmapped"
+
+    """
+    fgumi extract \
+        --inputs ${reads} \
+        --output ${prefix}.bam \
+        --threads ${task.cpus} \
+        --queue-memory ${queue_memory_mb} \
+        --queue-memory-per-thread \
+        ${args}
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.unmapped"
+    """
+    touch ${prefix}.bam
+    """
+}

modules/local/fgumi/filter/main.nf

@@ -0,0 +1,38 @@
+process FGUMI_FILTER {
+    tag "$meta.id"
+    label 'process_medium'
+
+    container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
+        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/95/954170443a820787c9e02ef2135ebb8ec29c6b03633b0d61b5fafa98c59a1cce/data'
+        : 'community.wave.seqera.io/library/fgumi_r-base_r-ggplot2_r-scales:09c99070b82c1c28'}"
+
+    input:
+    tuple val(meta), path(bam), path(fasta)
+
+    output:
+    tuple val(meta), path("${prefix}.filtered.bam"), emit: bam
+    tuple val(meta), path("${prefix}.filtering_metrics.txt"), optional: true, emit: filtering_metrics
+    tuple val("${task.process}"), val('fgumi'), eval("fgumi --version | sed 's/^fgumi //;q'"), topic: versions, emit: versions_fgumi
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.filter"
+
+    """
+    fgumi filter \
+        --input ${bam} \
+        --output ${prefix}.filtered.bam \
+        --ref ${fasta} \
+        ${args}
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.filter"
+    """
+    touch ${prefix}.filtered.bam
+    touch ${prefix}.filtering_metrics.txt
+    """
+}

modules/local/fgumi/group/main.nf

@@ -0,0 +1,44 @@
+process FGUMI_GROUP {
+    tag "$meta.id"
+    label 'process_medium'
+
+    container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
+        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/95/954170443a820787c9e02ef2135ebb8ec29c6b03633b0d61b5fafa98c59a1cce/data'
+        : 'community.wave.seqera.io/library/fgumi_r-base_r-ggplot2_r-scales:09c99070b82c1c28'}"
+
+    input:
+    tuple val(meta), path(bam)
+
+    output:
+    tuple val(meta), path("${prefix}.bam"), emit: bam
+    tuple val(meta), path("${prefix}.grouping_metrics.txt"), optional: true, emit: grouping_metrics
+    tuple val(meta), path("${prefix}.family_size_histogram.txt"), optional: true, emit: family_size_histogram
+    tuple val("${task.process}"), val('fgumi'), eval("fgumi --version | sed 's/^fgumi //;q'"), topic: versions, emit: versions_fgumi
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    // Derive per-thread queue memory from requested process resources.
+    def queue_memory_mb = (task.memory.mega / task.cpus * 0.75).intValue()
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.group"
+
+    """
+    fgumi group \
+        --input ${bam} \
+        --output ${prefix}.bam \
+        --threads ${task.cpus} \
+        --queue-memory ${queue_memory_mb} \
+        --queue-memory-per-thread \
+        ${args}
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.group"
+    """
+    touch ${prefix}.bam
+    touch ${prefix}.grouping_metrics.txt
+    touch ${prefix}.family_size_histogram.txt
+    """
+}

modules/local/fgumi/simplex/main.nf

@@ -0,0 +1,37 @@
+process FGUMI_SIMPLEX {
+    tag "$meta.id"
+    label 'process_medium'
+
+    container "${workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container
+        ? 'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/95/954170443a820787c9e02ef2135ebb8ec29c6b03633b0d61b5fafa98c59a1cce/data'
+        : 'community.wave.seqera.io/library/fgumi_r-base_r-ggplot2_r-scales:09c99070b82c1c28'}"
+
+    input:
+    tuple val(meta), path(bam)
+
+    output:
+    tuple val(meta), path("${prefix}.bam"), emit: bam
+    tuple val(meta), path("${prefix}.consensus_metrics.txt"), optional: true, emit: consensus_metrics
+    tuple val("${task.process}"), val('fgumi'), eval("fgumi --version | sed 's/^fgumi //;q'"), topic: versions, emit: versions_fgumi
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.simplex"
+
+    """
+    fgumi simplex \
+        --input ${bam} \
+        --output ${prefix}.bam \
+        ${args}
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}.fgumi.simplex"
+    """
+    touch ${prefix}.bam
+    touch ${prefix}.consensus_metrics.txt
+    """
+}