Merge pull request nf-core#12 from TalusBio/bb_feature/skip_prelim

Bb feature/skip prelim
ypriverol · Jan 10, 2024 · 3c80318 · 3c80318
2 parents e80ba58 + 639e507
commit 3c80318
Show file tree

Hide file tree

Showing 5 changed files with 77 additions and 30 deletions.
diff --git a/.gitignore b/.gitignore
@@ -18,3 +18,4 @@ debug_dir
 test_out
 
 lint_log.txt
+node_modules
diff --git a/modules/local/diann_preliminary_analysis/main.nf b/modules/local/diann_preliminary_analysis/main.nf
@@ -42,6 +42,8 @@ process DIANN_PRELIMINARY_ANALYSIS {
     # Precursor Tolerance unit was: ${meta['precursormasstoleranceunit']}
     # Fragment Tolerance unit was: ${meta['fragmentmasstoleranceunit']}
 
+    # Final mas accuracy is '${mass_acc}'
+
     diann   --lib ${predict_library} \\
             --f ${ms_file} \\
             --threads ${task.cpus} \\

diff --git a/nextflow.config b/nextflow.config
@@ -174,6 +174,10 @@ params {
     species_genes           = false
     diann_normalize         = true
     diann_speclib           = null
+
+    /// DIA-NN: Extras
+    skip_preliminary_analysis = false
+    empirical_assembly_log = null
 
     // MSstats general options
     msstats_remove_one_feat_prot    = true

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -940,6 +940,20 @@
                     "help_text": "If passed, will use that spectral library to carry out the DIA-NN search, instead of predicting one from the fasta file.",
                     "hidden": false
                 },
+                "skip_preliminary_analysis": {
+                    "type": "boolean",
+                    "description": "Skip the preliminary analysis step, thus use the passed spectral library as-is insted of generating a local concensus library.",
+                    "fa_icon": "fas fa-forward",
+                    "default": false,
+                    "hidden": false
+                },
+                "empirical_assembly_log": {
+                    "type": "string",
+                    "description": "The log file for the empirical assembly, Only used if `--skip_preliminary_analysis` is set to `true` and `--diann_speclib` is passed. If passed, will use that log file to carry out the DIA-NN search, instead of running a preliminary search.",
+                    "fa_icon": "fas fa-file",
+                    "help_text": "If passed, will use that log file to carry out the DIA-NN search, instead of predicting one from the fasta file.",
+                    "hidden": false
+                },
                 "diann_debug": {
                     "type": "integer",
                     "description": "Debug level",

diff --git a/workflows/dia.nf b/workflows/dia.nf
@@ -44,13 +44,14 @@ workflow DIA {
         result ->
         meta:   preprocessed_meta(result[0])
         ms_file:result[1]
-        }
+    }
         .set { ch_result }
 
     meta = ch_result.meta.unique { it[0] }
 
     DIANNCFG(meta)
-    ch_software_versions = ch_software_versions.mix(DIANNCFG.out.version.ifEmpty(null))
+    ch_software_versions = ch_software_versions
+        .mix(DIANNCFG.out.version.ifEmpty(null))
 
     //
     // MODULE: SILICOLIBRARYGENERATION
@@ -62,34 +63,48 @@ workflow DIA {
         speclib = SILICOLIBRARYGENERATION.out.predict_speclib
     }
 
-    //
-    // MODULE: DIANN_PRELIMINARY_ANALYSIS
-    //
-    DIANN_PRELIMINARY_ANALYSIS(ch_file_preparation_results.combine(speclib))
-    ch_software_versions = ch_software_versions.mix(DIANN_PRELIMINARY_ANALYSIS.out.version.ifEmpty(null))
+    if (params.skip_preliminary_analysis) {
+        assembly_log = Channel.fromPath(params.empirical_assembly_log)
+        empirical_library = Channel.fromPath(params.diann_speclib)
+        indiv_fin_analysis_in = ch_file_preparation_results.combine(ch_searchdb)
+            .combine(assembly_log)
+            .combine(empirical_library)
 
-    //
-    // MODULE: ASSEMBLE_EMPIRICAL_LIBRARY
-    //
-    // Order matters in DIANN, This shoudl be sorted for reproducible results.
-    ASSEMBLE_EMPIRICAL_LIBRARY(
-        ch_result.ms_file.collect(),
-        meta,
-        DIANN_PRELIMINARY_ANALYSIS.out.diann_quant.collect(),
-        speclib
-    )
-    ch_software_versions = ch_software_versions.mix(ASSEMBLE_EMPIRICAL_LIBRARY.out.version.ifEmpty(null))
+        empirical_lib = empirical_library
+    } else {
+        //
+        // MODULE: DIANN_PRELIMINARY_ANALYSIS
+        //
+        DIANN_PRELIMINARY_ANALYSIS(ch_file_preparation_results.combine(speclib))
+        ch_software_versions = ch_software_versions
+            .mix(DIANN_PRELIMINARY_ANALYSIS.out.version.ifEmpty(null))
+
+        //
+        // MODULE: ASSEMBLE_EMPIRICAL_LIBRARY
+        //
+        // Order matters in DIANN, This shoudl be sorted for reproducible results.
+        ASSEMBLE_EMPIRICAL_LIBRARY(
+            ch_result.ms_file.collect(),
+            meta,
+            DIANN_PRELIMINARY_ANALYSIS.out.diann_quant.collect(),
+            speclib
+        )
+        ch_software_versions = ch_software_versions
+            .mix(ASSEMBLE_EMPIRICAL_LIBRARY.out.version.ifEmpty(null))
+        indiv_fin_analysis_in = ch_file_preparation_results
+            .combine(ch_searchdb)
+            .combine(ASSEMBLE_EMPIRICAL_LIBRARY.out.log)
+            .combine(ASSEMBLE_EMPIRICAL_LIBRARY.out.empirical_library)
+
+        empirical_lib = ASSEMBLE_EMPIRICAL_LIBRARY.out.empirical_library
+    }
 
     //
     // MODULE: INDIVIDUAL_FINAL_ANALYSIS
     //
-    INDIVIDUAL_FINAL_ANALYSIS(
-        ch_file_preparation_results
-        .combine(ch_searchdb)
-        .combine(ASSEMBLE_EMPIRICAL_LIBRARY.out.log)
-        .combine(ASSEMBLE_EMPIRICAL_LIBRARY.out.empirical_library)
-    )
-    ch_software_versions = ch_software_versions.mix(INDIVIDUAL_FINAL_ANALYSIS.out.version.ifEmpty(null))
+    INDIVIDUAL_FINAL_ANALYSIS(indiv_fin_analysis_in)
+    ch_software_versions = ch_software_versions
+        .mix(INDIVIDUAL_FINAL_ANALYSIS.out.version.ifEmpty(null))
 
     //
     // MODULE: DIANNSUMMARY
@@ -103,9 +118,17 @@ workflow DIA {
         .ms_file.map { msfile -> file(msfile).getName() }
         .collect()
         .set { ms_file_names }
-    DIANNSUMMARY(ms_file_names, meta, ASSEMBLE_EMPIRICAL_LIBRARY.out.empirical_library,
-                    INDIVIDUAL_FINAL_ANALYSIS.out.diann_quant.collect(), ch_searchdb)
-    ch_software_versions = ch_software_versions.mix(DIANNSUMMARY.out.version.ifEmpty(null))
+
+    DIANNSUMMARY(
+        ms_file_names,
+        meta,
+        empirical_lib,
+        INDIVIDUAL_FINAL_ANALYSIS.out.diann_quant.collect(),
+        ch_searchdb)
+
+    ch_software_versions = ch_software_versions.mix(
+        DIANNSUMMARY.out.version.ifEmpty(null)
+    )
 
     //
     // MODULE: DIANNCONVERT
@@ -118,15 +141,18 @@ workflow DIA {
         ch_searchdb,
         DIANNSUMMARY.out.version
     )
-    ch_software_versions = ch_software_versions.mix(DIANNCONVERT.out.version.ifEmpty(null))
+    ch_software_versions = ch_software_versions
+        .mix(DIANNCONVERT.out.version.ifEmpty(null))
 
     //
     // MODULE: MSSTATS
     ch_msstats_out = Channel.empty()
     if (!params.skip_post_msstats) {
         MSSTATS(DIANNCONVERT.out.out_msstats)
         ch_msstats_out = MSSTATS.out.msstats_csv
-        ch_software_versions = ch_software_versions.mix(MSSTATS.out.version.ifEmpty(null))
+        ch_software_versions = ch_software_versions.mix(
+            MSSTATS.out.version.ifEmpty(null)
+        )
     }
 
     emit: