UMI Consensus read implementation

assets/schema_input.json

@@ -53,6 +53,26 @@
                 "description": "Run markdup in UMI-aware mode. This applies to Samtools only and requires the UMI to be in the read name.",
                 "default": false
             },
+            "umi_consensus": {
+                "meta": ["umi_consensus"],
+                "type": "boolean",
+                "description": "Enable UMI consensus workflow based on KAPA/fgbio processing for this sample.",
+                "default": false
+            },
+            "umi_strategy": {
+                "meta": ["umi_strategy"],
+                "type": "string",
+                "description": "UMI consensus strategy to apply when umi_consensus is enabled.",
+                "enum": ["kapa"],
+                "default": "kapa"
+            },
+            "umi_min_reads": {
+                "meta": ["umi_min_reads"],
+                "type": "integer",
+                "description": "Minimum reads per family for consensus calling when UMI consensus is enabled.",
+                "minimum": 1,
+                "default": 2
+            },
             "skip_trimming": {
                 "meta": ["skip_trimming"],
                 "type": "boolean",

assets/schema_sampleinfo.json

@@ -93,6 +93,26 @@
                 "description": "Run markdup in UMI-aware mode. This applies to Samtools only and requires the UMI to be in the read name.",
                 "default": false
             },
+            "umi_consensus": {
+                "meta": ["umi_consensus"],
+                "type": "boolean",
+                "description": "Enable UMI consensus workflow based on KAPA/fgbio processing for this sample.",
+                "default": false
+            },
+            "umi_strategy": {
+                "meta": ["umi_strategy"],
+                "type": "string",
+                "description": "UMI consensus strategy to apply when umi_consensus is enabled.",
+                "enum": ["kapa"],
+                "default": "kapa"
+            },
+            "umi_min_reads": {
+                "meta": ["umi_min_reads"],
+                "type": "integer",
+                "description": "Minimum reads per family for consensus calling when UMI consensus is enabled.",
+                "minimum": 1,
+                "default": 2
+            },
             "skip_trimming": {
                 "meta": ["skip_trimming"],
                 "type": "boolean",

conf/modules.config

@@ -85,6 +85,52 @@ process {
         ext.args2 = "--fast"
     }
 
+    // BWA aliases used by local wrappers
+    withName: '.*FASTQ_TO_CRAM:FASTQ_ALIGN_DNA:FASTQ_ALIGN_DNA_BWAMEM' {
+        cpus      = 16
+        memory    = 32.GB
+        ext.args  = {
+            [
+                "-K 100000000",
+                "-v 3",
+                "-Y",
+                "-c 250",
+                meta.readgroup ? "-R \"@RG\\t" + meta.readgroup.findResults { rg -> rg.value?.trim() ? "${rg.key}:${rg.value}" : null }.join("\\t") + "\"" : "",
+            ].join(" ").trim()
+        }
+        ext.args2 = "--fast"
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_CONSENSUS_KAPA:FASTQ_ALIGN_DNA_CONSENSUS:FASTQ_ALIGN_DNA_CONSENSUS_BWAMEM' {
+        cpus      = 16
+        memory    = 32.GB
+        ext.args  = {
+            [
+                "-K 100000000",
+                "-p",
+                "-v 3",
+                "-Y",
+                "-c 250",
+                meta.readgroup ? "-R \"@RG\\t" + meta.readgroup.findResults { rg -> rg.value?.trim() ? "${rg.key}:${rg.value}" : null }.join("\\t") + "\"" : "",
+            ].join(" ").trim()
+        }
+        ext.args2 = "--fast"
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_SAMTOOLS_SORT_PREP' {
+        cpus       = 16
+        memory     = 64.GB
+        ext.prefix = { "${meta.id}.umi_prep" }
+        ext.args   = '-l 1 --output-fmt bam'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_SAMTOOLS_MERGE_CRAM' {
+        cpus       = 16
+        memory     = 64.GB
+        ext.prefix = { "${meta.id}.umi_consensus_merged" }
+        ext.args   = '--output-fmt cram,version=3.0 --output-fmt-option archive'
+    }
+
     //// DRAGEN
     withName: '.*FASTQ_ALIGN_DNA:DRAGMAP_ALIGN' {
         cpus      = 16
@@ -231,6 +277,90 @@ process {
         }
     }
 
+    withName: '.*FASTQ_TO_CRAM:UMI_.*' {
+        cpus   = 16
+        memory = 64.GB
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_SAMTOOLS_PREP_TEMPLATE' {
+        ext.args   = '-F 256 -bh --output-fmt bam'
+        ext.args2  = '-O -u --output-fmt bam'
+        ext.args3  = '-m --output-fmt bam'
+        ext.args4  = '--template-coordinate --output-fmt bam'
+        ext.prefix = { "${meta.id}.template" }
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_FGBIO_GROUPREADSBYUMI' {
+        ext.args = '--edits 1 -t RX'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_FGBIO_CALLMOLECULARCONSENSUSREADS' {
+        cpus   = 16
+        memory = 64.GB
+        ext.args = '--error-rate-pre-umi 45 --error-rate-post-umi 40 --max-reads 50 --output-per-base-tags false --read-name-prefix consensus'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_FGBIO_FILTERCONSENSUSREADS' {
+        cpus   = 16
+        memory = 64.GB
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_FGBIO_ZIPPERBAMS' {
+        ext.args2 = '--tags-to-reverse Consensus --tags-to-revcomp Consensus'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_SAMTOOLS_SORT_FINAL' {
+        ext.args = '-l 9 --output-fmt bam'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_SORT_ZIP_MAPPED' {
+        ext.prefix = { "${meta.id}.consensus_mapped_qname" }
+        ext.args   = '-n -l 1 --no-PG --output-fmt bam'
+    }
+    withName: '.*UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_SORT_ZIP_MAPPED' {
+        ext.prefix = { "${meta.id}.consensus_mapped_qname" }
+        ext.args   = '-n -l 1 --no-PG --output-fmt bam'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_SORT_ZIP_UNMAPPED' {
+        ext.prefix = { "${meta.id}.consensus_unmapped_qname" }
+        ext.args   = '-n -l 1 --no-PG --output-fmt bam'
+    }
+    withName: '.*UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_SORT_ZIP_UNMAPPED' {
+        ext.prefix = { "${meta.id}.consensus_unmapped_qname" }
+        ext.args   = '-n -l 1 --no-PG --output-fmt bam'
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_STRIP_PG' {
+        ext.prefix = { "${meta.id}.consensus_mapped_qname_nopg" }
+    }
+    withName: '.*UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_STRIP_PG$' {
+        ext.prefix = { "${meta.id}.consensus_mapped_qname_nopg" }
+    }
+
+    withName: '.*FASTQ_TO_CRAM:UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_STRIP_PG_UNMAPPED' {
+        ext.prefix = { "${meta.id}.consensus_unmapped_qname_nopg" }
+    }
+    withName: '.*UMI_CONSENSUS_KAPA:UMI_SAMTOOLS_STRIP_PG_UNMAPPED' {
+        ext.prefix = { "${meta.id}.consensus_unmapped_qname_nopg" }
+    }
+
+    withName: '.*FASTQ_ALIGN_DNA_CONSENSUS:FASTQ_ALIGN_DNA_CONSENSUS_SNAP' {
+        ext.args = {
+            [
+                '-pairedInterleavedFastq',
+                '-b-',
+                '-sm 20',
+                '-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()
+        }
+    }
+
     // coverage
     //// Mosdepth
     withName: '.*COVERAGE:MOSDEPTH' {

docs/README.md

@@ -3,7 +3,7 @@
 The nf-cmgg/preprocessing documentation is split into the following pages:
 
 - [Usage](usage.md)
-  - An overview of how the pipeline works, how to run it and a description of all of the different command-line flags.
+  - An overview of how the pipeline works, how to run it, command-line flags, and UMI consensus implementation details.
 - [Output](output.md)
   - An overview of the different results produced by the pipeline and how to interpret them.
 - [Parameters](parameters.md)

docs/images/metro_map_dark.md

@@ -29,7 +29,9 @@ graph TD
     FALCO -->|qc| MULTIQC_LIBRARY
     FASTP -->|qc| MULTIQC_LIBRARY
 
+    FASTP -->|main| UMI_CONSENSUS
     FASTP -->|main| ALIGN
+    UMI_CONSENSUS -->|main| ALIGN
     ALIGN -->|main| MARKDUP
     MARKDUP -->|main| CRAM_OUT