Merge pull request #130 from Joon-Klaps/add-mmseqs-cluster-option

Update new variable mmseqs_cluster_mode default 0
Joon-Klaps · Jun 24, 2024 · d52fb0a · d52fb0a
2 parents 9038504 + 40cc340
commit d52fb0a
Show file tree

Hide file tree

Showing 4 changed files with 13 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -12,6 +12,7 @@ Initial release of Joon-Klaps/viralgenie, created with the [nf-core](https://nf-
 - Set default umitools dedup strategy to cluster ([#126](https://github.com/Joon-Klaps/viralgenie/pull/126))
 - Include sspace for contig extension ([#123](https://github.com/Joon-Klaps/viralgenie/pull/123))
 - Include both krakenreport &nodes.dmp in taxonomy ([#128](https://github.com/Joon-Klaps/viralgenie/pull/128))
+- Update new variable mmseqs_cluster_mode default 0 ([#130](https://github.com/Joon-Klaps/viralgenie/pull/130))
 
 ### `Fixed`
 

diff --git a/conf/modules.config b/conf/modules.config
@@ -608,7 +608,7 @@ process {
                         "--min-seq-id ${params.identity_threshold}",    // List as a match if above for clustering
                         "-c 0.700",                                     // fraction aligned of the shorter sequence
                         "--cov-mode 2",                                 // coverage of query
-                        "--cluster-mode 1",                             // connected components ~ BLASTclust
+                        "--cluster-mode ${params.mmseqs_cluster_mode}", // connected components ~ BLASTclust
                     ].join(' ').trim()
                 publishDir =[
                     [
@@ -629,7 +629,7 @@ process {
                         "-c 0.700",                                     // fraction aligned of the shorter sequence
                         "--cov-mode 2",                                 // coverage of query
                         "-s 7",                                         // senstive
-                        "--cluster-mode 1",                             // connected components ~ BLASTclust
+                        "--cluster-mode ${params.mmseqs_cluster_mode}", // connected components ~ BLASTclust
                     ].join(' ').trim()
                 publishDir =[
                     [

diff --git a/nextflow.config b/nextflow.config
@@ -91,6 +91,7 @@ params {
     // > Clustering
     cluster_method                   = 'mash'
     network_clustering               = 'connected_components'
+    mmseqs_cluster_mode              = 0
     identity_threshold               = 0.6
     min_contig_size                  = 500
     max_n_perc                       = 50

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -446,6 +446,15 @@
                     "enum": ["cdhitest", "vsearch", "mmseqs-linclust", "mmseqs-cluster", "vrhyme", "mash", "diamond"],
                     "fa_icon": "fas fa-bezier-curve"
                 },
+                "mmseqs_cluster_mode": {
+                    "type": "integer",
+                    "default": 0,
+                    "fa_icon": "fas fa-project-diagram",
+                    "description": "Specify the algorithm to partition the network graph from mmseqs",
+                    "help_text": "The Greedy Set cover (0) algorithm is an approximation for the NP-complete optimization problem called set cover.\nConnected component (1) uses transitive connection to cover more remote homologs.\nGreedy incremental (2) works analogous to CD-HIT clustering algorithm.",
+                    "minimum": 0,
+                    "maximum": 3
+                },
                 "network_clustering": {
                     "type": "string",
                     "default": "connected_components",