Merge pull request #14 from cidgoh/asm_QC

Update assembly QC parameter descriptions
cidgoh · Jul 11, 2023 · ee8ddb2 · ee8ddb2
2 parents ea1459d + 8f6f49c
commit ee8ddb2
Show file tree

Hide file tree

Showing 6 changed files with 29 additions and 25 deletions.
diff --git a/conf/modules.config b/conf/modules.config
@@ -141,13 +141,13 @@ process {
             mode: params.publish_dir_mode,
             pattern: '*.{gfa,log,fasta,yml,corrections,LastGraph}'
         ]
-        ext.args = "--assembler ${params.assembler} \
+        ext.args = "--assembler ${params.sr_assembler}"
                     --gsize ${params.assembly_genomesize} \
                     --minlen ${params.min_contig_len} \
                     --mincov ${params.min_contig_coverage}"
     }
 
-    withName: RENAME_SHOVILL {
+    withName: RENAME_CTG {
         publishDir = [
             path: { "${params.outdir}/assembly/${meta.id}" },
             mode: params.publish_dir_mode,

diff --git a/modules/local/rename_shovill/main.nf → modules/local/rename_ctg/main.nf b/modules/local/rename_shovill/main.nf → modules/local/rename_ctg/main.nf
@@ -1,4 +1,4 @@
-process RENAME_SHOVILL {
+process RENAME_CTG {
     tag "$meta.id"
     label 'process_low'
 

diff --git a/nextflow.config b/nextflow.config
@@ -103,9 +103,9 @@ params {
     // Aligner used for dehosting ('minimap2' or 'bwa')
     dehosting_aligner          = 'minimap2'
     // Path to (human) reference genome used for dehosting
-    reference_genome           = "/mnt/cidgoh-object-storage/database/reference_genomes/human/GRCh38.p14/GCF_000001405.40/GCF_000001405.40_GRCh38.p14_genomic.fna"
+    host_genome           = "/mnt/cidgoh-object-storage/database/reference_genomes/human/GRCh38.p14/GCF_000001405.40/GCF_000001405.40_GRCh38.p14_genomic.fna"
     // ID/name for reference genome 
-    ref_genome_id              = 'test'
+    host_genome_id              = 'test'
 
     // bwa options
     // Whether to sort ('sort') or view ('view') the BAM file obtained from BWA
@@ -138,18 +138,21 @@ params {
     show_hidden_params         = false
     schema_ignore_params       = 'genomes'
 
+    // Shovill options
+    sr_assembler               = 'spades' // WGS_ASSEMBLY: choice of genome assembler for short read data [skesa,spades,megahit,velvet]
+
     // CheckM options
-    checkm_db                  = false
+    checkm_db                  = false // ASSEMBLY_QC: path to local CheckM lineages directory
 
     // QUAST options
-    reference                  = false
-    reference_gff              = false
-    combine_quast              = false
+    reference_genome_fasta            = false // ASSEMBLY_QC: path to reference genome file 
+    reference_genome_gff              = false // ASSEMBLY_QC: path to reference genome annotation file (.GFF)
+    combine_quast              = false // ASSEMBLY_QC: whether to aggregate individual QUAST results file
 
     // BUSCO options
-    busco_lineage              = "auto"
-    busco_lineages_path        = "/mnt/cidgoh-object-storage/database/busco"
-    busco_config               = false
+    busco_lineage              = "auto" // ASSEMBLY_QC: the BUSCO lineage to use, or "auto" to automatically select lineage
+    busco_lineages_path        = "/mnt/cidgoh-object-storage/database/busco" // ASSEMBLY_QC: path to local BUSCO lineages directory
+    busco_config               = false // ASSEMBLY_QC: path to BUSCO config file
 
     // Config options
     custom_config_version      = 'master'

diff --git a/subworkflows/local/assembly_qc.nf b/subworkflows/local/assembly_qc.nf
@@ -24,18 +24,18 @@ workflow ASSEMBLY_QC {
     if ( params.combine_quast ) {
         QUAST(
             assembly.map{it[1]}.collect(),
-            params.reference ? params.reference : [],
-            params.reference_gff ? params.reference_gff : [],
-            params.reference ? assembly.map{ true } : params.reference,
-            params.reference_gff ? assembly.map{ true } : params.reference_gff
+            params.reference_genome_fasta ? params.reference_genome_fasta : assembly.map{ [] },
+            params.reference_genome_gff ? params.reference_genome_gff : assembly.map{ [] },
+            params.reference_genome_fasta ? assembly.map{ true } : params.reference_genome_fasta,
+            params.reference_genome_gff ? assembly.map{ true } : params.reference_genome_gff
         )
     } else {
         QUAST(
             assembly.map{it[1]},
-            params.reference ? params.reference : assembly.map{ [] },
-            params.reference_gff ? params.reference_gff : assembly.map{ [] },
-            params.reference ? assembly.map{ true } : params.reference,
-            params.reference_gff ? assembly.map{ true } : params.reference_gff
+            params.reference_genome_fasta ? params.reference_genome_fasta : assembly.map{ [] },
+            params.reference_genome_gff ? params.reference_genome_gff : assembly.map{ [] },
+            params.reference_genome_fasta ? assembly.map{ true } : params.reference_genome_fasta,
+            params.reference_genome_gff ? assembly.map{ true } : params.reference_genome_gff
         )
     }
 

diff --git a/subworkflows/local/wgs_assembly.nf b/subworkflows/local/wgs_assembly.nf
@@ -1,6 +1,6 @@
 // import modules
 include { SHOVILL } from '../../modules/nf-core/shovill'
-include { RENAME_SHOVILL } from '../../modules/local/rename_shovill'
+include { RENAME_CTG as RENAME_CTG_SHOVILL; RENAME_CTG as RENAME_CTG_DRAGONFLYE } from '../../modules/local/rename_ctg'
 // include {  } from '../../modules/local/rename_shovill'
 include { DRAGONFLYE } from '../../modules/nf-core/dragonflye'
 
@@ -11,8 +11,9 @@ workflow WGS_ASSEMBLY {
     main:
     if (params.mode=="nanopore"){
         reads | DRAGONFLYE
+        RENAME_CTG_DRAGONFLYE(DRAGONFLYE.out.contigs, reads.map{ "fa" })
 
-        contigs = DRAGONFLYE.out.contigs
+        contigs = RENAME_CTG_DRAGONFLYE.out
         corrections = []
         assembly_log = DRAGONFLYE.out.log
         raw_contigs = DRAGONFLYE.out.raw_contigs
@@ -24,9 +25,9 @@ workflow WGS_ASSEMBLY {
         reads.filter{ it[0].single_end == false } | SHOVILL
 
         // rename contig files
-        RENAME_SHOVILL(SHOVILL.out.contigs, reads.map{ "fa" })
+        RENAME_CTG_SHOVILL(SHOVILL.out.contigs, reads.map{ "fa" })
 
-        contigs = RENAME_SHOVILL.out
+        contigs = RENAME_CTG_SHOVILL.out
         corrections = SHOVILL.out.corrections
         assembly_log = SHOVILL.out.log
         raw_contigs = SHOVILL.out.raw_contigs

diff --git a/workflows/seqqc.nf b/workflows/seqqc.nf
@@ -116,7 +116,7 @@ workflow SEQQC {
             if(!params.skip_taxonomy_qc){
             TAXONOMY_QC (
                 ch_reads_qc,
-                params.reference_genome
+                params.host_genome
                 )
                 ch_assembly_reads = TAXONOMY_QC.out.ch_tax_qc_reads
             }