Make GTDBTK explicitly optional, and more flexible with input databases

nf-core · Jul 29, 2023 · 1b829d6 · 1b829d6
1 parent 4f3a554
commit 1b829d6
Show file tree

Hide file tree

Showing 5 changed files with 15 additions and 15 deletions.
diff --git a/modules/local/gtdbtk_db_preparation.nf b/modules/local/gtdbtk_db_preparation.nf
@@ -10,7 +10,7 @@ process GTDBTK_DB_PREPARATION {
     path(database)
 
     output:
-    tuple val("${database.toString().replace(".tar.gz", "")}"), path("database/*")
+    tuple val("${database.toString().replace(".tar.gz", "")}"), path("database/*"), emit: db
 
     script:
     """

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -518,7 +518,7 @@
                 },
                 "gtdb_db": {
                     "type": "string",
-                    "description": "Specify the location of a GTDBTK database. Can be either an uncompressed directory or a `.tar.gz` archive. If not specified will be downloaded for you.",
+                    "description": "Specify the location of a GTDBTK database. Can be either an uncompressed directory or a `.tar.gz` archive. If not specified will be downloaded for you when GTDBTK or binning QC is not skipped.",
                     "default": "https://data.ace.uq.edu.au/public/gtdb/data/releases/release202/202.0/auxillary_files/gtdbtk_r202_data.tar.gz"
                 },
                 "gtdbtk_min_completeness": {

diff --git a/subworkflows/local/binning.nf b/subworkflows/local/binning.nf
@@ -130,9 +130,9 @@ workflow BINNING {
     ch_versions = ch_versions.mix(GUNZIP_UNBINS.out.versions.first())
 
     emit:
-    bins                                         = ch_binning_results_gunzipped.dump(tag: "ch_binning_results_gunzipped")
+    bins                                         = ch_binning_results_gunzipped
     bins_gz                                      = ch_binning_results_gzipped_final
-    unbinned                                     = ch_splitfasta_results_gunzipped.dump(tag: "ch_splitfasta_results_gunzipped")
+    unbinned                                     = ch_splitfasta_results_gunzipped
     unbinned_gz                                  = SPLIT_FASTA.out.unbinned
     metabat2depths                               = METABAT2_JGISUMMARIZEBAMCONTIGDEPTHS.out.depth
     versions                                     = ch_versions

diff --git a/subworkflows/local/gtdbtk.nf b/subworkflows/local/gtdbtk.nf
@@ -59,12 +59,13 @@ workflow GTDBTK {
                 return [it[0], it[1]]
         }
 
-    if ( gtdb.extension == '.gz' ) {
+    if ( gtdb.extension == 'gz' ) {
         // Expects to be tar.gz!
-        ch_db_for_gtdbtk = GTDBTK_DB_PREPARATION ( gtdb ).out
+        ch_db_for_gtdbtk = GTDBTK_DB_PREPARATION ( gtdb ).db
     } else if ( gtdb.isDirectory() ) {
         // Make up meta id to match expected channel cardinality for GTDBTK
-        ch_db_for_gtdbtk = gtdb
+        ch_db_for_gtdbtk = Channel
+                            .of(gtdb)
                             .map{
                                 [ it.toString().split('/').last(), it ]
                             }
@@ -74,8 +75,8 @@ workflow GTDBTK {
     }
 
     GTDBTK_CLASSIFYWF (
-        ch_filtered_bins.passed.groupTuple().dump(tag: "BINS"),
-        ch_db_for_gtdbtk.dump(tag: "SHOULD_BE_GTDB_DB")
+        ch_filtered_bins.passed.groupTuple(),
+        ch_db_for_gtdbtk
     )
 
     GTDBTK_SUMMARY (

diff --git a/workflows/mag.nf b/workflows/mag.nf
@@ -208,10 +208,9 @@ if (params.genomad_db){
 gtdb = ( params.skip_binqc || params.skip_gtdbtk ) ? false : params.gtdb_db
 
 if (gtdb) {
-    ch_gtdb = Channel
-        .value(file( "${gtdb}", checkIfExists: true))
+    gtdb = file( "${gtdb}", checkIfExists: true)
 } else {
-    ch_gtdb = Channel.empty()
+    gtdb = []
 }
 
 if(params.metaeuk_db && !params.skip_metaeuk) {
@@ -720,12 +719,12 @@ workflow MAG {
 
 
         } else {
-            ch_binning_results_bins = BINNING.out.bins.dump(tag: 'BINNING.out.bins')
+            ch_binning_results_bins = BINNING.out.bins
                 .map { meta, bins ->
                     def meta_new = meta + [domain: 'unclassified']
                     [meta_new, bins]
                 }
-            ch_binning_results_unbins = BINNING.out.unbinned.dump(tag: 'BINNING.out.unbins')
+            ch_binning_results_unbins = BINNING.out.unbinned
                 .map { meta, bins ->
                     def meta_new = meta + [domain: 'unclassified']
                     [meta_new, bins]
@@ -892,7 +891,7 @@ workflow MAG {
                     ch_gtdb_bins,
                     ch_busco_summary,
                     ch_checkm_summary,
-                    ch_gtdb
+                    gtdb
                 )
                 ch_versions = ch_versions.mix(GTDBTK.out.versions.first())
                 ch_gtdbtk_summary = GTDBTK.out.summary