Merge pull request #121 from nf-core/mysolid-bugs

Fix bugs pointed out by @mysolid on slack
nf-core · May 22, 2024 · 74dfdf6 · 74dfdf6
2 parents 5ddc061 + e273f70
commit 74dfdf6
Show file tree

Hide file tree

Showing 5 changed files with 12 additions and 89 deletions.
diff --git a/bin/annotation.py b/bin/annotation.py
@@ -43,7 +43,7 @@
 df_intergenic['transcript_id'] = 'intergenic_' + df_intergenic['name']
 
 # Convert attributes to a dictionary
-df['attributes'] = df['attributes'].apply(lambda row: dict([[value.strip(r'"') for value in entry.strip().split(' ')] for entry in row.split(';') if entry]))
+df['attributes'] = df['attributes'].apply(lambda row: dict([[value.strip(r'"') for value in entry.strip().split(' ', 1)] for entry in row.split(';') if entry]))
 # Keep only the attributes we want
 df['attributes'] = df['attributes'].apply(lambda row: {key: row[key] for key in attributes if key in row})
 # Convert attributes to columns

diff --git a/bin/backsplice_gen.sh b/bin/backsplice_gen.sh
diff --git a/conf/modules.config b/conf/modules.config
@@ -590,13 +590,14 @@ if (!params.skip_trimming) {
         ext.suffix = "tidy.bed"
     }
 
-    withName: FASTA {
-        ext.when = { params.module.split(',').contains('circrna_discovery') }
+    withName: ADD_BACKSPLICE {
+        ext.args = "'{ if (/^>/) { print \$0 } else { start = substr(\$0, 1, 25); print \$0 start } }'"
+        ext.suffix = "backspliced.fa"
         publishDir = [
             path: { "${params.outdir}/circrna_discovery/${meta.tool}/${meta.id}" },
             mode: params.publish_dir_mode,
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
-            pattern: "*.fasta"
+            pattern: "*.backspliced.fa"
         ]
     }
 

diff --git a/modules/local/fasta/main.nf b/modules/local/fasta/main.nf
diff --git a/subworkflows/local/circrna_discovery.nf b/subworkflows/local/circrna_discovery.nf
@@ -27,7 +27,6 @@ include { SJDB as DCC_MATE2_SJDB                         } from '../../modules/l
 include { DCC                                            } from '../../modules/local/dcc/dcc'
 include { DCC_FILTER                                     } from '../../modules/local/dcc/filter'
 include { MAPSPLICE_ALIGN                                } from '../../modules/local/mapsplice/align'
-include { FASTA                                          } from '../../modules/local/fasta'
 include { MERGE_TOOLS                                    } from '../../modules/local/count_matrix/merge_tools'
 include { COUNTS_COMBINED                                } from '../../modules/local/count_matrix/combined'
 include { CIRCEXPLORER2_REFERENCE as CIRCEXPLORER2_REF   } from '../../modules/local/circexplorer2/reference'
@@ -40,6 +39,8 @@ include { CIRCEXPLORER2_ANNOTATE as MAPSPLICE_ANNOTATE   } from '../../modules/n
 include { CIRCEXPLORER2_FILTER as MAPSPLICE_FILTER       } from '../../modules/local/circexplorer2/filter'
 include { UPSET as UPSET_SAMPLES                         } from '../../modules/local/upset'
 include { UPSET as UPSET_ALL                             } from '../../modules/local/upset'
+include { BEDTOOLS_GETFASTA                              } from '../../modules/nf-core/bedtools/getfasta'
+include { GAWK as ADD_BACKSPLICE                         } from '../../modules/nf-core/gawk'
 
 workflow CIRCRNA_DISCOVERY {
 
@@ -272,13 +273,15 @@ workflow CIRCRNA_DISCOVERY {
     // FASTA WORKFLOW:
     //
 
-    FASTA( ch_annotation_bed_merged, fasta )
+    BEDTOOLS_GETFASTA( ch_annotation_bed_merged, fasta )
+    ADD_BACKSPLICE( BEDTOOLS_GETFASTA.out.fasta, [])
 
-    ch_versions = ch_versions.mix(FASTA.out.versions)
+    ch_versions = ch_versions.mix(BEDTOOLS_GETFASTA.out.versions)
+    ch_versions = ch_versions.mix(ADD_BACKSPLICE.out.versions)
 
     emit:
     circrna_bed12 = ch_annotation_bed_merged
-    fasta = FASTA.out.analysis_fasta
+    fasta = ADD_BACKSPLICE.out.output
     annotation_bed = REMOVE_SCORE_STRAND.out.output
     annotation_gtf = COMBINE_ANNOTATION_GTFS.out.sorted
     counts_bed = COUNTS_COMBINED.out.counts_bed