updated config defaults for various organisms (#105)

* updated config defaults for various organisms * updated changelog
phac-nml · Sep 4, 2024 · 2976de7 · 2976de7
1 parent 056d7f8
commit 2976de7
Show file tree

Hide file tree

Showing 2 changed files with 80 additions and 7 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -22,10 +22,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fixed typo in metagenomic QC message. See [PR 103](https://github.com/phac-nml/mikrokondo/pull/103)
 
 - Fixed spelling issues issues in config values. See [PR 95](https://github.com/phac-nml/mikrokondo/pull/95)
+
 - Fixed the headers specified in the nextflow.config file for Kraken2. See [PR 96](https://github.com/phac-nml/mikrokondo/pull/96)
 
 ### `Added`
 
+- Added additional organism QC parameters to defaults. See [PR 105](https://github.com/phac-nml/mikrokondo/pull/105)
+
 - Updated locidex to version 0.2.3. See [PR 96](https://github.com/phac-nml/mikrokondo/pull/96)
 
 - Added module for automatic selection of locidex databases through configuration of a locidex database collection. See [PR 96](https://github.com/phac-nml/mikrokondo/pull/96)

diff --git a/nextflow.config b/nextflow.config
@@ -839,27 +839,97 @@ params {
             search = "Klebsiella"
             raw_average_quality = 30
             min_n50 = 100000
-            max_n50 = 6000000
+            max_n50 = 6500000
             min_nr_contigs = 1
             max_nr_contigs = 500
-            fixed_genome_size = 6000000
             min_length = 4500000
-            max_length = 6000000
+            max_length = 6500000
             max_checkm_contamination = 3.0
-            min_average_coverage = 30
+            min_average_coverage = 50
         }
         staphylococcus {
             search = "Staphylococcus"
             raw_average_quality = 30
+            min_n50 = 30000
+            max_n50 = 3500000
+            min_nr_contigs = 1
+            max_nr_contigs = 250
+            min_length = 2500000
+            max_length = 3500000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
+        }
+        enterobacter {
+            search = "Enterobacter"
+            raw_average_quality = 30
+            min_n50 = 100000
+            max_n50 = 6500000
+            min_nr_contigs = 1
+            max_nr_contigs = 500
+            min_length = 4500000
+            max_length = 6500000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
+        }
+        citrobacter {
+            search = "Citrobacter"
+            raw_average_quality = 30
+            min_n50 = 100000
+            max_n50 = 6500000
+            min_nr_contigs = 1
+            max_nr_contigs = 500
+            min_length = 4500000
+            max_length = 6500000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
+        }
+        acinetobacter {
+            search = "Acinetobacter"
+            raw_average_quality = 30
+            min_n50 = 100000
+            max_n50 = 6500000
+            min_nr_contigs = 1
+            max_nr_contigs = 500
+            min_length = 4500000
+            max_length = 6500000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
+        }
+        pseudomonas {
+            search = "Pseudomonas"
+            raw_average_quality = 30
             min_n50 = 100000
+            max_n50 = 7000000
+            min_nr_contigs = 1
+            max_nr_contigs = 500
+            min_length = 5000000
+            max_length = 7000000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
+        }
+        enterococcus {
+            search = "Enterococcus"
+            raw_average_quality = 30
+            min_n50 = 15000
             max_n50 = 3500000
             min_nr_contigs = 1
             max_nr_contigs = 550
-            fixed_genome_size = 3500000
-            min_length = 2000000
+            min_length = 2500000
             max_length = 3500000
             max_checkm_contamination = 3.0
-            min_average_coverage = 30
+            min_average_coverage = 50
+        }
+            clostridioides {
+            search = "Clostridioides"
+            raw_average_quality = 30
+            min_n50 = 20000
+            max_n50 = 4500000
+            min_nr_contigs = 1
+            max_nr_contigs = 600
+            min_length = 4000000
+            max_length = 4500000
+            max_checkm_contamination = 3.0
+            min_average_coverage = 50
         }
         fallthrough {
             search = "No organism specific QC data available."