Improvements to Salmon SA and mapping stats

nf-core · Nov 22, 2024 · 9f0e3b0 · 9f0e3b0
1 parent 085cd4a
commit 9f0e3b0
Show file tree

Hide file tree

Showing 14 changed files with 480 additions and 281 deletions.
diff --git a/conf/modules.config b/conf/modules.config
@@ -26,42 +26,38 @@ process {
         ]
     }
 
-    withName: 'COMBINE_FILES|CREATE_TRANSCRIPTOME_FASTA|CREATE_TRANSCRIPTOME_FASTA_GFFREAD|REPLACE_ATTRIBUTE_GFF_STAR_SALMON*' {
-
-        publishDir       = [
-            [
-                path: { "${params.outdir}/references" },
-                mode: params.publish_dir_mode,
+    withName: 'COMBINE_FILES|CREATE_TRANSCRIPTOME_FASTA|CREATE_TRANSCRIPTOME_FASTA_GFFREAD' {
+        publishDir = [
+            [path: { "${params.outdir}/references" },
+             mode: params.publish_dir_mode,
             ]
         ]
     }
 
+    withName: 'REPLACE_ATTRIBUTE_GFF_STAR_SALMON_HOST|REPLACE_ATTRIBUTE_GFF_STAR_SALMON_PATHOGEN' {
+         publishDir = [ enabled: false ]
+    }
 
-    withName: 'UNCOMPRESS_HOST_FASTA_GENOME|UNCOMPRESS_PATHOGEN_FASTA_GENOME|UNCOMPRESS_HOST_GFF|UNCOMPRESS_HOST_TRNA_GFF|UNCOMPRESS_PATHOGEN_GFF' {
 
-        publishDir       = [
-            [
-                path: { "${params.outdir}/references" },
-                mode: params.publish_dir_mode,
+    withName: 'UNCOMPRESS_HOST_FASTA_GENOME|UNCOMPRESS_PATHOGEN_FASTA_GENOME|UNCOMPRESS_HOST_GFF|UNCOMPRESS_PATHOGEN_GFF|COMBINE_FILES_PATHOGEN_HOST_GFF' {
+        publishDir = [
+            [path: { "${params.outdir}/references" },
+             mode: params.publish_dir_mode,
             ]
         ]
-
         ext.prefix = 'uncompressed'
     }
 
 
     withName: 'COMBINE_HOST_GFF_FILES' {
-
-        publishDir       = [
-            [
-                path: { "${params.outdir}/references/htseq" },
-                mode: params.publish_dir_mode,
+        publishDir = [
+            [path: { "${params.outdir}/references/htseq" },
+             mode: params.publish_dir_mode,
             ]
         ]
     }
 
-    withName: 'REPLACE_ATTRIBUTE_GFF_STAR_SALMON_PATHOGEN|REPLACE_ATTRIBUTE_GFF_STAR_SALMON_HOST|REPLACE_ATTRIBUTE_GFF_STAR_SALMON_TRNA_FILE|COMBINE_HOST_GENOME_TRNA_GFF_STAR_SALMON|REPLACE_GENE_FEATURE_GFF_HOST_SALMON|REPLACE_GENE_FEATURE_GFF_PATHOGEN_SALMON|COMBINE_FILES_PATHOGEN_HOST_GFF|EXTRACT_ANNOTATIONS_PATHOGEN_SALMON|EXTRACT_ANNOTATIONS_HOST_SALMON' {
-
+    withName: 'REPLACE_GENE_FEATURE_GFF_HOST_SALMON|REPLACE_GENE_FEATURE_GFF_PATHOGEN_SALMON|EXTRACT_ANNOTATIONS_PATHOGEN_SALMON|EXTRACT_ANNOTATIONS_HOST_SALMON' {
         publishDir       = [
             [
                 path: { "${params.outdir}/references/salmon" },
@@ -73,8 +69,6 @@ process {
 
 
 
-
-
     withName: FASTQC {
         ext.args = '--quiet'
         ext.when         = { !(params.skip_tools && params.skip_tools.split(',').contains('fastqc')) }
@@ -85,7 +79,6 @@ process {
                 pattern: "*{html,zip}"
             ]
         ]
-        container = 'quay.io/biocontainers/fastqc:0.11.9--0'
     }
 
     withName: FASTQC_AFTER_TRIMMING {
@@ -156,7 +149,9 @@ process {
         ]
     }
 
-    withName: 'NFCORE_DUALRNASEQ:DUALRNASEQ:SALMON_SELECTIVE_ALIGNMENT:EXTRACT_PROCESSED_READS' {
+
+
+    withName: 'NFCORE_DUALRNASEQ:DUALRNASEQ:SALMON_SELECTIVE_ALIGNMENT:COMBINE_PROCESSED_READS|EXTRACT_PROCESSED_READS|COLLATE_PROCESSED_READS' {
         publishDir = [
             path: { "${params.outdir}/mapping_statistics/salmon_SA/" },
             mode: params.publish_dir_mode
@@ -226,12 +221,6 @@ process {
         ]
     }
 
-    withName: 'NFCORE_DUALRNASEQ:DUALRNASEQ:SALMON_ALIGNMENT_BASED:EXTRACT_PROCESSED_READS' {
-        publishDir = [
-            path: { "${params.outdir}/mapping_statistics/STAR_SALMON/" },
-            mode: params.publish_dir_mode
-        ]
-    }
 
     withName: 'NFCORE_DUALRNASEQ:DUALRNASEQ:SALMON_ALIGNMENT_BASED:TXIMPORT' {
         publishDir = [

diff --git a/conf/test_hackathon.config b/conf/test_hackathon.config
@@ -25,17 +25,17 @@ params {
     // Genome references
 
     // fasta_host  = "https://github.com/nf-core/test-datasets/raw/dualrnaseq/references/GRCh38.p13_sub.fasta"
-    fasta_host  = "data/GRCh38.p13_sub.fasta"
-    gff_host = "data/Human_gencode.v33_sub.gff3"
+    host_fasta_genome  = "data/GRCh38.p13_sub.fasta"
+    host_gff = "data/Human_gencode.v33_sub.gff3"
     // gff_host = "https://github.com/nf-core/test-datasets/raw/dualrnaseq/references/Human_gencode.v33_sub.gff3"
-    gff_host_tRNA = 'data/human.tRNAs.gff'
-    transcript_fasta_pathogen = "data/SL1344_sub_transcriptome.fasta"
-    transcript_fasta_host = "data/Human_gencode.v33_sub_transcriptome.fasta"
+    // gff_host_tRNA = 'data/human.tRNAs.gff'
+    pathogen_fasta_transcripts = "data/SL1344_sub_transcriptome.fasta"
+    host_fasta_transcripts = "data/Human_gencode.v33_sub_transcriptome.fasta"
     libtype = ""
 
     // fasta_pathogen  = "https://github.com/nf-core/test-datasets/raw/dualrnaseq/references/SL1344_sub.fasta"
     // gff_pathogen = "https://github.com/nf-core/test-datasets/raw/dualrnaseq/references/SL1344_sub.gff3"
-    fasta_pathogen  = "data/SL1344_sub.fasta"
-    gff_pathogen = "data/SL1344_sub.gff3"
+    pathogen_fasta_genome  = "data/SL1344_sub.fasta"
+    pathogen_gff = "data/SL1344_sub.gff3"
 
 }
diff --git a/lib/WorkflowDualrnaseq.groovy b/lib/WorkflowDualrnaseq.groovy
@@ -12,11 +12,11 @@ class WorkflowDualrnaseq {
     public static void initialise(params, log) {
         genomeExistsError(params, log)
 
-
-        if (!params.fasta_host) {
-            log.error "Host genome fasta file not specified with e.g. '--fasta_host genome.fa' or via a detectable config file."
-            System.exit(1)
-        }
+        // TODO - could add in checks for host and pathogen fasta and gff files??
+        // if (!params.fasta_host) {
+        //     log.error "Host genome fasta file not specified with e.g. '--fasta_host genome.fa' or via a detectable config file."
+        //     System.exit(1)
+        // }
     }
 
     //

diff --git a/modules/local/collate_processed_reads.nf b/modules/local/collate_processed_reads.nf
@@ -0,0 +1,17 @@
+process COLLATE_PROCESSED_READS {
+    label 'process_high'
+    conda "python=3.8.3"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ? 'nfcore/dualrnaseq:dev' : 'nfcore/dualrnaseq:dev' }"
+
+    input:
+    file partial_results
+
+    output:
+    path 'total_processed_reads.tsv'
+
+    script:
+    """
+    # Concatenate all partial result files into the master file
+     cat ${partial_results.join(' ')} > total_processed_reads.tsv
+    """
+}
diff --git a/modules/local/extract_annotations.nf b/modules/local/extract_annotations.nf
@@ -23,7 +23,7 @@ process EXTRACT_ANNOTATIONS {
 
     script:
     def args = task.ext.args ?: ''
-    def prefix = task.ext.prefix ?: "extract_annotations_${organism}_${quantifier}"
+    def prefix = task.ext.prefix ?: "extracted_annotations_${organism}_${quantifier}"
 
     """
     python $workflow.projectDir/bin/extract_annotations_from_gff.py  \\

diff --git a/modules/local/extract_processed_reads.nf b/modules/local/extract_processed_reads.nf
@@ -1,5 +1,6 @@
+#!/usr/bin/env nextflow
 process EXTRACT_PROCESSED_READS {
-    tag "extract_process_reads_${process}"
+    tag "extract_processed_reads_${process}"
     label 'process_high'
 
     conda "python=3.8.3"
@@ -29,4 +30,3 @@ process EXTRACT_PROCESSED_READS {
     fi
     """
 }
-
diff --git a/modules/local/replace_attribute.nf b/modules/local/replace_attribute.nf
@@ -1,7 +1,6 @@
 // ONE FOR HOST_GENOME
-// ONE FOR HOST_TRNA
 process REPLACE_ATTRIBUTE_GFF_STAR_SALMON {
-	tag "repl_attribute_host_tRNA_gff"
+	tag "repl_GFF_attributes"
 
 	label 'process_high'
 

diff --git a/modules/local/test.nf b/modules/local/test.nf
diff --git a/nextflow.config b/nextflow.config
@@ -40,24 +40,29 @@ params {
     // A default run will include the following software:
     // Fastqc, Cutadapt, Fastqc (after trimming), Salmon SA, mapping stats
     // To add or remove software, change the following flags/params
-    run_salmon_selective_alignment = true
-    run_salmon_alignment_based_mode = false //true
     skip_tools                 = null
 
-
+//fasta_host = null
     // --------------
     // Reference files
     // --------------
 
+    // Names of the organisms
+    host_organism = 'host' // Change to custom name if desired, ie Human_hela_cells
+    pathogen_organism = 'pathogen' // Change to custom name if desired, ie Salmonella_SL1344
+
     // Host
-    fasta_host                 = null
-    transcript_fasta_host      = null
+    host_fasta_genome           = null
+    host_fasta_transcripts      = null
+    host_gff                    = null
 
     // Pathogen
-    transcript_fasta_pathogen  = null
+    pathogen_fasta_genome       = null
+    pathogen_fasta_transcripts  = null
+    pathogen_gff                = null
 
     // Misc
-    genome                     = null
+    //genome                     = null
 
 
     // --------------
@@ -69,15 +74,14 @@ params {
     gene_feature_gff_to_quantify_host               = ["exon", "tRNA"]
     extract_annotations_host_salmon_feature         = 'exon' //'quant'
     extract_annotations_host_salmon_attribute       = 'Parent'  // currently it is required to pass value with capital letter
-    host_organism = 'host' // Change to custom name if desired, ie Human_hela_cells
+
 
     // Pathogen
     pathogen_gff_attribute     = 'locus_tag'
     gene_feature_gff_to_quantify_pathogen = ["gene", "sRNA", "tRNA", "rRNA"]
-    pathogen_organism = 'pathogen' // Change to custom name if desired, ie Salmonella_SL1344
+
 
     // Misc
-    htseq_quantifier = 'quant'
     // read_transcriptome_fasta_host_from_file = false
     // read_transcriptome_fasta_pathogen_from_file = false
 
@@ -87,12 +91,15 @@ params {
     // --------------
     igenomes_base              = 's3://ngi-igenomes/igenomes'
     igenomes_ignore            = false
+    igenome = null
+    ifasta = null
 
 
     // --------------
     // Software options
     // -------------- 
 
+    mapping_stats = true
     // Fastqc
     fastqc_args = null
 
@@ -108,11 +115,21 @@ params {
 
 
     // Salmon selective alignment
+    run_salmon_SA = true
     salmon_sa_index_args = '-k 21'
     salmon_sa_args = '--softclipOverhangs'
 
+    // Salmon alignment based mode
+    run_salmon_AB = false
     salmon_ab_args = null
 
+    // STAR genome alignment
+    run_star = false
+
+    // Run HTSeq
+    run_htseq = false
+    htseq_quantifier = 'quant'
+
 
 
     // --------------