Implement parameter for skipping rank_genes_groups

nf-core · Sep 22, 2024 · 8981320 · 8981320
1 parent 31210ac
commit 8981320
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 4 deletions.
diff --git a/nextflow.config b/nextflow.config
@@ -30,6 +30,7 @@ params {
     clustering_resolutions       = '0.5,1.0'
     cluster_per_label            = false
     cluster_global               = true
+    skip_rankgenesgroups         = false
     celltypist_model             = ''
     cellbender_epochs            = 150
     var_aggr_method              = 'mean'

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -152,6 +152,11 @@
                     "default": true,
                     "description": "Create a global UMAP and clustering (for each integration method)"
                 },
+                "skip_rankgenesgroups": {
+                    "type": "boolean",
+                    "default": false,
+                    "description": "Skip the rank_genes_groups step. For large datasets, the pipeline might fail due to high memory usage in this step. Use this option to skip it."
+                },
                 "celltypist_model": {
                     "type": "string",
                     "default": "",

diff --git a/subworkflows/local/cluster.nf b/subworkflows/local/cluster.nf
@@ -69,10 +69,12 @@ workflow CLUSTER {
     ch_uns = ch_uns.mix(SCANPY_PAGA.out.uns)
     ch_multiqc_files = ch_multiqc_files.mix(SCANPY_PAGA.out.multiqc_files)
 
-    SCANPY_RANKGENESGROUPS(SCANPY_LEIDEN.out.h5ad)
-    ch_versions = ch_versions.mix(SCANPY_RANKGENESGROUPS.out.versions)
-    ch_uns = ch_uns.mix(SCANPY_RANKGENESGROUPS.out.uns)
-    ch_multiqc_files = ch_multiqc_files.mix(SCANPY_RANKGENESGROUPS.out.multiqc_files)
+    if (!params.skip_rankgenesgroups) {
+        SCANPY_RANKGENESGROUPS(SCANPY_LEIDEN.out.h5ad)
+        ch_versions = ch_versions.mix(SCANPY_RANKGENESGROUPS.out.versions)
+        ch_uns = ch_uns.mix(SCANPY_RANKGENESGROUPS.out.uns)
+        ch_multiqc_files = ch_multiqc_files.mix(SCANPY_RANKGENESGROUPS.out.multiqc_files)
+    }
 
     emit:
     obs = ch_obs