Merge pull request #400 from TalusBio/fix/random-sample

Make files selected with `random_preanalysis` reproducible.

CHANGELOG.md

@@ -9,6 +9,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - [#386](https://github.com/bigbio/quantms/pull/386) Make validation of ontology terms optional
 
+### Fixed
+
+- [#400](https://github.com/bigbio/quantms/pull/400) The random file selection when using `random_preanalysis` with DIANN is now reproducible.
+
 ### `Changed`
 
 ### `Fixed`

nextflow.config

@@ -201,6 +201,7 @@ params {
     skip_preliminary_analysis    = false
     empirical_assembly_log       = null
     random_preanalysis           = false
+    random_preanalysis_seed      = 42
     empirical_assembly_ms_n      = 200
 
 

nextflow_schema.json

@@ -1091,6 +1091,12 @@
                     "default": false,
                     "fa_icon": "far fa-check-square"
                 },
+                "random_preanalysis_seed": {
+                    "type": "integer",
+                    "description": "Set the random seed for the random selection of spectrum files to generate the empirical library.",
+                    "default": 42,
+                    "fa_icon": false
+                },
                 "empirical_assembly_ms_n": {
                     "type": "integer",
                     "description": "The number of randomly selected spectrum files.",

workflows/dia.nf

@@ -76,9 +76,10 @@ workflow DIA {
         // MODULE: DIANN_PRELIMINARY_ANALYSIS
         //
         if (params.random_preanalysis) {
-            preanalysis_seed = 2024
             preanalysis_subset = ch_file_preparation_results
-                .randomSample(params.empirical_assembly_ms_n, preanalysis_seed)
+                .toSortedList()
+                .flatMap()
+                .randomSample(params.empirical_assembly_ms_n, params.random_preanalysis_seed)
             empirical_lib_files = preanalysis_subset
                 .map { result -> result[1] }
                 .collect()