Add Kallisto quantification while removing deprecated option to use d…

…edup reads for fusion calls
mskcc · Jun 28, 2023 · b968173 · b968173
1 parent 1ec8d9c
commit b968173
Show file tree

Hide file tree

Showing 13 changed files with 304 additions and 48 deletions.
diff --git a/conf/igenomes.config b/conf/igenomes.config
@@ -14,6 +14,7 @@ params {
             fasta          = "${params.igenomes_base}/Homo_sapiens/GATK/GRCh37/Sequence/WholeGenomeFasta/human_g1k_v37_decoy.fasta"
             gtf            = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Annotation/Genes/genes.gtf"
             refflat        = "${params.igenomes_base}/Homo_sapiens/Ensembl/GRCh37/Annotation/Genes/refFlat.txt.gz"
+            cdna           = "https://ftp.ensembl.org/pub/release-75/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh37.75.cdna.all.fa.gz"
             starfusion_url = "https://data.broadinstitute.org/Trinity/CTAT_RESOURCE_LIB/__genome_libs_StarFv1.10/GRCh37_gencode_v19_CTAT_lib_Mar012021.plug-n-play.tar.gz"
             arriba_blacklist       = "/usr/local/var/lib/arriba/blacklist_hg19_hs37d5_GRCh37_v2.3.0.tsv.gz"
             arriba_known_fusions   = "/usr/local/var/lib/arriba/known_fusions_hg19_hs37d5_GRCh37_v2.3.0.tsv.gz"
@@ -41,6 +42,7 @@ params {
             arriba_blacklist       = "/usr/local/var/lib/arriba/blacklist_hg38_GRCh38_v2.3.0.tsv.gz"
             arriba_known_fusions   = "/usr/local/var/lib/arriba/known_fusions_hg38_GRCh38_v2.3.0.tsv.gz"
             arriba_protein_domains = "/usr/local/var/lib/arriba/protein_domains_hg38_GRCh38_v2.3.0.gff3"
+            cdna           = "https://ftp.ensembl.org/pub/release-86/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh38.cdna.all.fa.gz"
         }
         'smallGRCh37' {
             fasta          = "https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/homo_sapiens/genome/genome.fasta"
@@ -50,6 +52,7 @@ params {
             arriba_blacklist       = "/usr/local/var/lib/arriba/blacklist_hg19_hs37d5_GRCh37_v2.3.0.tsv.gz"
             arriba_known_fusions   = "/usr/local/var/lib/arriba/known_fusions_hg19_hs37d5_GRCh37_v2.3.0.tsv.gz"
             arriba_protein_domains = "/usr/local/var/lib/arriba/protein_domains_hg19_hs37d5_GRCh37_v2.3.0.gff3"
+            cdna           = "ftp://ftp.ensembl.org/pub/grch37/current/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh37.75.cdna.abinitio.fa.gz"
         }
 /*
         'hg38' {

diff --git a/conf/modules.config b/conf/modules.config
@@ -344,6 +344,25 @@ process {
         time     = { check_max( 20.h * task.attempt, 'time' ) }
     }
 
+    withName: KALLISTO_QUANT {
+        ext.args = {
+            [
+                "--plain-text",
+                "--bias",
+                "-b 100",
+                meta.strand == "forward" ?
+                    "--fr-stranded" :
+                    (
+                        meta.strand == "reverse" ?
+                        "--rf-stranded" :
+                        ""
+                    )
+            ].join(" ")
+        }
+        ext.fragment_len = params.kallisto_fragment_len
+        ext.sd = params.kallisto_fragment_sd
+    }
+
     withName: ARRIBA {
         ext.args = {
             [

diff --git a/modules.json b/modules.json
@@ -45,6 +45,16 @@
                         "git_sha": "c8e35eb2055c099720a75538d1b8adb3fb5a464c",
                         "installed_by": ["modules"]
                     },
+                    "kallisto/index": {
+                        "branch": "master",
+                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
+                        "installed_by": ["modules"]
+                    },
+                    "kallisto/quant": {
+                        "branch": "master",
+                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
+                        "installed_by": ["modules"]
+                    },
                     "multiqc": {
                         "branch": "master",
                         "git_sha": "ee80d14721e76e2e079103b8dcd5d57129e584ba",

diff --git a/modules/nf-core/kallisto/index/main.nf b/modules/nf-core/kallisto/index/main.nf
diff --git a/modules/nf-core/kallisto/index/meta.yml b/modules/nf-core/kallisto/index/meta.yml
diff --git a/modules/nf-core/kallisto/quant/main.nf b/modules/nf-core/kallisto/quant/main.nf
diff --git a/modules/nf-core/kallisto/quant/meta.yml b/modules/nf-core/kallisto/quant/meta.yml
diff --git a/nextflow.config b/nextflow.config
@@ -29,7 +29,6 @@ params {
     save_align_intermeds       = false
 
     // Fusion
-    dedup_umi_for_fusions      = false
     run_oncokb_fusionannotator = false
     cosmic_usr                 = null
     fusion_report_cutoff       = 1
@@ -39,6 +38,11 @@ params {
     // rseqc_modules can include ['bam_stat','inner_distance','infer_experiment','junction_annotation','junction_saturation','read_distribution','read_duplication','tin']
     rseqc_modules              = ['bam_stat','inner_distance','infer_experiment','junction_annotation','junction_saturation','read_distribution','read_duplication']
 
+    // Quantification
+    dedup_umi_for_kallisto     = true
+    kallisto_fragment_len      = 500
+    kallisto_fragment_sd       = 150
+
     // MultiQC options
     multiqc_config             = null
     multiqc_title              = null

diff --git a/subworkflows/local/extract_dedup_fq.nf b/subworkflows/local/extract_dedup_fq.nf
@@ -0,0 +1,24 @@
+include { SAMTOOLS_BAM2FQ } from '../../modules/nf-core/samtools/bam2fq/main'
+
+workflow EXTRACT_DEDUP_FQ {
+    take:
+    bam
+
+    main:
+    ch_versions = Channel.empty()
+
+    SAMTOOLS_BAM2FQ(
+        bam,
+        true
+    )
+    ch_versions = ch_versions.mix(SAMTOOLS_BAM2FQ.out.versions.first())
+
+    dedup_reads = SAMTOOLS_BAM2FQ.out.reads
+        .map{ meta, reads ->
+            [meta, reads.findAll{ !(it.getName().endsWith("singleton.fq.gz") || it.getName().endsWith("other.fq.gz")) }]
+        }
+
+    emit:
+    dedup_reads  = dedup_reads
+    ch_versions  = ch_versions
+}
diff --git a/subworkflows/local/merge_reads.nf b/subworkflows/local/merge_reads.nf
diff --git a/subworkflows/local/prepare_references.nf b/subworkflows/local/prepare_references.nf
@@ -9,6 +9,8 @@ include { GUNZIP                         } from '../../modules/nf-core/gunzip/ma
 include { STARFUSION_DOWNLOAD            } from '../../modules/local/starfusion/download/main'
 include { FUSIONCATCHER_DOWNLOAD         } from '../../modules/local/fusioncatcher/download/main'
 include { FUSIONREPORT_DOWNLOAD          } from '../../modules/local/fusionreport/download/main'
+include { KALLISTO_INDEX                 } from '../../modules/local/kallisto/index/main'
+
 
 workflow PREPARE_REFERENCES {
 
@@ -70,6 +72,9 @@ workflow PREPARE_REFERENCES {
     //cosmic_passwd = params.cosmic_passwd ?: ""
     FUSIONREPORT_DOWNLOAD()
 
+    KALLISTO_INDEX(params.cdna)
+    ch_versions = ch_versions.mix(KALLISTO_INDEX.out.versions)
+
     emit:
     star_index         = star_index
     // Convert queue channel to value channel so it never gets poison pilled
@@ -84,6 +89,7 @@ workflow PREPARE_REFERENCES {
     fusioncatcher_ref  = fusioncatcher_ref
     fusion_report_db   = FUSIONREPORT_DOWNLOAD.out.reference
     rseqc_bed          = UCSC_GENEPREDTOBED.out.bed.map{it[1]}.first()
+    kallisto_index     = KALLISTO_INDEX.out.idx
     ch_versions        = ch_versions
 
 }