Merge pull request #63 from nf-core/better_fimo_combine_results

Better fimo combine results

Author: nictru

conf/modules.config

@@ -12,12 +12,6 @@
 
 process {
 
-    publishDir = [
-        path: { "${params.outdir}/all/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" },
-        mode: params.publish_dir_mode,
-        saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
-    ]
-
     /*
     PREPARE_GENOME
     */
@@ -589,12 +583,40 @@ process {
         ]
     }
 
-    withName: ".*:FIMO:COMBINE_RESULTS" {
+    withName: ".*:FIMO:CONCAT_FILTER_GFF" {
+        ext.args = "'/^#/ { next } NF==0 { next } { print }'"
+        publishDir = [
+            // Unsorted intermediate file, disable publishing
+            enabled: false
+        ]
+    }
+
+    withName: ".*:FIMO:SORT_GFF" {
+        ext.args = "-t \$'\\t' -k1,1V -k4,4n -k5,5n"
+        ext.prefix = {"${meta.id}_sorted"}
+        publishDir = [
+            path: { "${params.outdir}/06_fimo/04_combined_results/gff" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: ".*:FIMO:CONCAT_FILTER_TSV" {
+        ext.args = "-v HEADER='motif_id\\tmotif_alt_id\\tsequence_name\\tstart\\tstop\\tstrand\\tscore\\tp-value\\tq-value\\tmatched_sequence' 'BEGIN {print HEADER} FNR==1 {next} /^#/ {next} NF==0 {next} {print}'"
+        publishDir = [
+            // Unsorted intermediate file, disable publishing
+            enabled: false
+        ]
+    }
+
+    withName: ".*:FIMO:SORT_TSV" {
+        ext.args = "-k motif_id:N -k sequence_name:N -k start:n -k stop:n -k strand -k score:nr -k p-value:n -k q-value:n -k matched_sequence"
         publishDir = [
-            path: { "${params.outdir}/06_fimo/04_combined_results" },
+            path: { "${params.outdir}/06_fimo/04_combined_results/tsv" },
             mode: params.publish_dir_mode,
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
         ]
+        memory = { 36.GB * task.attempt }
     }
 
     /*

docs/output.md

@@ -315,7 +315,7 @@ FIMO analysis is particularly valuable for users who need detailed binding site
       - `*.tsv`, `*.gff`, `*.html`, `*.xml`, `*cisml.xml`, `*best_site.narrowPeak`.
   - `04_combined_results/`
     - `<condition>_<assay>.tsv`
-    - `<condition>_<assay>.gff`: Collated FIMO hits across inputs (COMBINE_RESULTS).
+    - `<condition>_<assay>.gff`: Collated FIMO hits across inputs.
 
 </details>
 

modules.json

@@ -46,6 +46,11 @@
                         "git_sha": "41dfa3f7c0ffabb96a6a813fe321c6d1cc5b6e46",
                         "installed_by": ["modules"]
                     },
+                    "csvtk/sort": {
+                        "branch": "master",
+                        "git_sha": "d01263b60ce84d777f09819709e94dd17b11ce92",
+                        "installed_by": ["modules"]
+                    },
                     "deseq2/differential": {
                         "branch": "master",
                         "git_sha": "2e24bf8b6b045e1cf56a2907306b008113a686d0",