moved validation parameter section from logratioanalysis.nf to mainwo…

…rkflow.nf
nf-core · Oct 17, 2023 · c205add · c205add
1 parent 1bd95cf
commit c205add
Show file tree

Hide file tree

Showing 5 changed files with 95 additions and 50 deletions.
diff --git a/conf/test.config b/conf/test.config
@@ -23,7 +23,6 @@ params {
     max_time   = '6.h'
 
     // Input data
-
     input = 'https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/mus_musculus/rnaseq_expression/SRP254919.samplesheet.csv'
     matrix =  'https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/mus_musculus/rnaseq_expression/SRP254919.salmon.merged.gene_counts.top1000cov.tsv'
     contrasts =  'https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/mus_musculus/rnaseq_expression/SRP254919.contrasts.csv'
@@ -32,19 +31,29 @@ params {
     gtf = 'https://ftp.ensembl.org/pub/release-81/gtf/mus_musculus/Mus_musculus.GRCm38.81.gtf.gz'
 
     // Observations
-    observations_id_col        = 'sample'
-    observations_name_col      = 'sample'
+    observations_id_col       = 'sample'
+    observations_name_col     = 'sample'
 
     // Apply a higher filter to check that the filtering works
-    filtering_min_abundance=10
+    filtering_min_abundance   = 10
+
+    /* 
+    differential abundance specific parameters 
+    */
 
     // Exploratory
-    exploratory_main_variable      = 'contrasts'
+    exploratory_main_variable = 'contrasts'
 
     // Test dataset is too small for the nsub default value
-    deseq2_vst_nsub                   = 500
+    deseq2_vst_nsub           = 500
 
     // Activate GSEA
     gsea_run = true
     gsea_gene_sets = 'https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/mus_musculus/gene_set_analysis/mh.all.v2022.1.Mm.symbols.gmt'
+
+    /* 
+    logratio analysis specific parameters 
+    */
+
+    // TODO add parameters
 }
diff --git a/conf/test_affy.config b/conf/test_affy.config
@@ -35,6 +35,10 @@ params {
     // Exploratory
     exploratory_main_variable      = 'contrasts'
 
+    /* 
+    differential abundance specific parameters 
+    */
+
     // Differential
     differential_max_pval            = 0.05
     differential_max_qval            = 0.05
@@ -43,4 +47,10 @@ params {
     // Activate GSEA
     gsea_run = true
     gsea_gene_sets = 'https://raw.githubusercontent.com/nf-core/test-datasets/differentialabundance/testdata/h.all.v2022.1.Hs.symbols.gmt'
+
+    /* 
+    logratio analysis specific parameters 
+    */
+
+    // TODO add parameters
 }
diff --git a/main.nf b/main.nf
@@ -37,6 +37,8 @@ if (params.help) {
 }
 
 // Validate input parameters
+// this will check that the given parameters met the ones declared in json files
+// it will also check all the required inputs are fullfilled
 if (params.validate_params) {
     validateParameters()
 }

diff --git a/subworkflows/logratioanalysis.nf b/subworkflows/logratioanalysis.nf
@@ -4,44 +4,6 @@
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 */
 
-include { paramsSummaryLog; paramsSummaryMap } from 'plugin/nf-validation'
-
-def logo = NfcoreTemplate.logo(workflow, params.monochrome_logs)
-def citation = '\n' + WorkflowMain.citation(workflow) + '\n'
-def summary_params = paramsSummaryMap(workflow)
-
-// Print parameter summary log to screen
-log.info logo + paramsSummaryLog(workflow) + citation
-
-// Validate input parameters
-// TODO make it specific for logratioanalysis || or maybe it common for the whole workflow
-WorkflowDifferentialabundance.initialise(params,log)
-
-// Check input path parameters to see if they exist
-def checkPathParamList = [ 
-    params.input,
-    params.matrix,
-    params.gtf
-]
-for (param in checkPathParamList) { if (param) { file(param, checkIfExists: true) } }
-
-// Define meta
-def exp_meta = [ "id": params.study_name  ]
-
-// Check input samplesheet
-if (params.input) { 
-    input_file = file(params.input, checkIfExists: true)
-    ch_input = Channel.of([ exp_meta, input_file ]) 
-    } else { 
-        exit 1, 'Input samplesheet not specified!' 
-    }
-// Check input matrix
-if (params.matrix) { 
-        matrix_file = file(params.matrix, checkIfExists: true)
-        ch_in_raw = Channel.of([ exp_meta, matrix_file])
-    } else { 
-        error("Input matrix not specified!")
-    }
 
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -75,11 +37,14 @@ include { PROPR_PROPD as PROPR_DIFFERENTIALPROPORTIONALITY  } from '../modules/n
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 */
 
-// Info required for completion email and summary
-def multiqc_report = []
-
 workflow LOGRATIOANALYSIS {
 
+    take:
+    ch_input
+    ch_in_raw
+
+
+    main:
 
     /*
      * DATA HANDLING
@@ -103,6 +68,8 @@ workflow LOGRATIOANALYSIS {
 
     // TODO need to handle different data structures here (eg. affy, etc)
 
+    // TODO use contrast
+
     // TODO currently the zeros can be directly handled in the downstream analysis by replacing them with the min value
     // TODO add zero imputation methods
 

diff --git a/workflows/mainworkflow.nf b/workflows/mainworkflow.nf
@@ -1,17 +1,72 @@
+/*
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+    PRINT PARAMS SUMMARY
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+*/
+
+include { paramsSummaryLog; paramsSummaryMap } from 'plugin/nf-validation'
+
+def logo = NfcoreTemplate.logo(workflow, params.monochrome_logs)
+def citation = '\n' + WorkflowMain.citation(workflow) + '\n'
+def summary_params = paramsSummaryMap(workflow)
+
+log.info logo + paramsSummaryLog(workflow) + citation
+
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     VALIDATE INPUTS
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 */
 
-// TODO move the validation block from the subworkflows to here
+// Validate input parameters
+WorkflowDifferentialabundance.initialise(params, log)
+
+// Check input path parameters to see if they exist
+def checkPathParamList = [ params.input ]
+for (param in checkPathParamList) { if (param) { file(param, checkIfExists: true) } }
+
+// Check mandatory parameters
+def exp_meta = [ "id": params.study_name  ]
+if (params.input) { 
+    ch_input = Channel.of([ exp_meta, file(params.input, checkIfExists: true) ]) 
+} else { 
+    exit 1, 'Input samplesheet not specified!' 
+}
+
+// handle different data formats
+if (params.study_type == 'affy_array'){
+
+    if (params.affy_cel_files_archive) {
+        ch_celfiles = Channel.of([ exp_meta, file(params.affy_cel_files_archive, checkIfExists: true) ])
+    } else {
+        error("CEL files archive not specified!")
+    }
+
+} else {
+    // If this is not microarray data or maxquant output, and this an RNA-seq dataset,
+    // then assume we're reading from a matrix
+
+    if (params.study_type == "rnaseq" && params.matrix) {
+        matrix_file = file(params.matrix, checkIfExists: true)
+        ch_in_raw = Channel.of([ exp_meta, matrix_file])
+    } else {
+        error("Input matrix not specified!")
+    }
+
+}
+
+report_file = file(params.report_file, checkIfExists: true)
+logo_file = file(params.logo_file, checkIfExists: true)
+css_file = file(params.css_file, checkIfExists: true)
+citations_file = file(params.citations_file, checkIfExists: true)
 
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     CONFIG FILES
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 */
 
+// TODO maybe create two config files depending on running logratioanalysis or differentialabundance
 
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -33,7 +88,6 @@ if (params.run_partial_correlation || params.run_proportionality || params.run_d
 */
 
 
-
 /*
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     RUN MAIN WORKFLOW
@@ -47,7 +101,10 @@ workflow MAINWORKFLOW {
     }
 
     if (params.run_partial_correlation || params.run_proportionality || params.run_differential_proportionality){
-        LOGRATIOANALYSIS()
+        LOGRATIOANALYSIS(
+            ch_input,
+            ch_in_raw
+        )
     }
 
 }