Merge pull request #9 from ajmaurais/consolidate_qc

Consolidate QC command line output
mriffle · Aug 1, 2024 · d15743b · d15743b
2 parents 3acdc02 + 1c8df96
commit d15743b
Show file tree

Hide file tree

Showing 5 changed files with 44 additions and 88 deletions.
diff --git a/container_images.config b/container_images.config
@@ -6,7 +6,7 @@ params {
         panorama_client:       'quay.io/protio/panorama-client:1.1.0',
         encyclopedia:          'quay.io/protio/encyclopedia:2.12.30-2',
         encyclopedia3_mriffle: 'quay.io/protio/encyclopedia:3.0.0-MRIFFLE',
-        qc_pipeline:           'quay.io/mauraisa/dia_qc_report:2.0.0',
+        qc_pipeline:           'quay.io/mauraisa/dia_qc_report:2.2.2',
         proteowizard:          'quay.io/protio/pwiz-skyline-i-agree-to-the-vendor-licenses:3.0.24172-63d00b1'
     ]
 }
diff --git a/modules/encyclopedia.nf b/modules/encyclopedia.nf
@@ -28,7 +28,7 @@ process ENCYCLOPEDIA_SEARCH_FILE {
         path("${mzml_file}.features.txt"), emit: features
         path("${mzml_file}.encyclopedia.txt"), emit: results_targets
         path("${mzml_file}.encyclopedia.decoy.txt"), emit: results_decoys
-        
+
 
     script:
     """

diff --git a/modules/qc_report.nf b/modules/qc_report.nf
@@ -22,13 +22,13 @@ process MAKE_EMPTY_FILE {
         path("${file_name}")
 
     script:
-        """
-        touch ${file_name}
-        """
+    """
+    touch ${file_name}
+    """
 }
 
 process PARSE_REPORTS {
-    publishDir "${params.result_dir}/qc_report", pattern: '*.db3', enabled: params.qc_report.normalization_method == null, failOnError: true, mode: 'copy'
+    publishDir "${params.result_dir}/qc_report", pattern: '*.db3', failOnError: true, mode: 'copy'
     publishDir "${params.result_dir}/qc_report", pattern: '*.stdout', failOnError: true, mode: 'copy'
     publishDir "${params.result_dir}/qc_report", pattern: '*.stderr', failOnError: true, mode: 'copy'
     label 'process_high_memory'
@@ -41,81 +41,46 @@ process PARSE_REPORTS {
 
     output:
         path('*.db3'), emit: qc_report_db
+        path('*.qmd'), emit: qc_report_qmd
         path("*.stdout"), emit: stdout
         path("*.stderr"), emit: stderr
 
     script:
-        def metadata_arg = replicate_metadata.name == 'EMPTY' ? '' : "-m $replicate_metadata"
+    def metadata_arg = replicate_metadata.name == 'EMPTY' ? '' : "-m $replicate_metadata"
+
+    if(params.qc_report.normalization_method == null)
         """
         dia_qc parse --ofname qc_report_data.db3 ${metadata_arg} \
             --groupBy ${params.skyline.group_by_gene ? 'gene' : 'protein'} \
             '${replicate_report}' '${precursor_report}' \
             > >(tee "parse_data.stdout") 2> >(tee "parse_data.stderr")
-        """
 
-    stub:
-        """
-        touch stub.stdout stub.stderr stub.db3 stub.qmd
+        dia_qc qc_qmd ${format_flags(params.qc_report.standard_proteins, '--addStdProtein')} \
+            ${format_flags(params.qc_report.color_vars, '--addColorVar')} \
+            qc_report_data.db3 \
+            > >(tee "make_qmd.stdout") 2> >(tee "make_qmd.stderr")
         """
-}
 
-process NORMALIZE_DB {
-    publishDir "${params.result_dir}/qc_report", pattern: '*.db3', failOnError: true, mode: 'copy'
-    publishDir "${params.result_dir}/qc_report", pattern: '*.stdout', failOnError: true, mode: 'copy'
-    publishDir "${params.result_dir}/qc_report", pattern: '*.stderr', failOnError: true, mode: 'copy'
-    label 'process_high_memory'
-    container params.images.qc_pipeline
-
-    input:
-        path qc_report_db
-
-    output:
-        path('*_normalized.db3'), emit: qc_report_db
-        path("*.stdout"), emit: stdout
-        path("*.stderr"), emit: stderr
-
-    script:
+    else
         """
-        # It is necissary to make a copy to avoid breaking nextflow's caching
-        cp ${qc_report_db} ${qc_report_db.baseName}_normalized.db3
+        dia_qc parse --ofname qc_report_data.db3 ${metadata_arg} \
+            --groupBy ${params.skyline.group_by_gene ? 'gene' : 'protein'} \
+            '${replicate_report}' '${precursor_report}' \
+            > >(tee "parse_data.stdout") 2> >(tee "parse_data.stderr")
 
-        dia_qc normalize -m=${params.qc_report.normalization_method} ${qc_report_db.baseName}_normalized.db3 \
+        dia_qc normalize -m=${params.qc_report.normalization_method} qc_report_data.db3 \
             > >(tee "normalize_db.stdout") 2> >(tee "normalize_db.stderr" >&2)
-        """
-
-    stub:
-        """
-        touch stub.stdout stub.stderr stub_normalized.db3
-        """
-}
 
-process GENERATE_QC_QMD {
-    publishDir "${params.result_dir}/qc_report", pattern: '*.qmd', failOnError: true, mode: 'copy'
-    publishDir "${params.result_dir}/qc_report", pattern: '*.stdout', failOnError: true, mode: 'copy'
-    publishDir "${params.result_dir}/qc_report", pattern: '*.stderr', failOnError: true, mode: 'copy'
-    label 'process_high'
-    container params.images.qc_pipeline
-
-    input:
-        path qc_report_db
-
-    output:
-        path('*.qmd'), emit: qc_report_qmd
-        path("*.stdout"), emit: stdout
-        path("*.stderr"), emit: stderr
-
-    script:
-        """
         dia_qc qc_qmd ${format_flags(params.qc_report.standard_proteins, '--addStdProtein')} \
             ${format_flags(params.qc_report.color_vars, '--addColorVar')} \
-            ${qc_report_db} \
+            qc_report_data.db3 \
             > >(tee "make_qmd.stdout") 2> >(tee "make_qmd.stderr")
         """
 
     stub:
-        """
-        touch stub.stdout stub.stderr stub.qmd
-        """
+    """
+    touch stub.stdout stub.stderr stub.db3 stub.qmd
+    """
 }
 
 process EXPORT_TABLES {
@@ -134,15 +99,15 @@ process EXPORT_TABLES {
         path("*.stderr"), emit: stderr
 
     script:
-        """
-        dia_qc db_export --precursorTables=30 --proteinTables=30 ${precursor_db} \
-            > >(tee "export_tables.stdout") 2> >(tee "export_tables.stderr")
-        """
+    """
+    dia_qc db_export --precursorTables=30 --proteinTables=30 ${precursor_db} \
+        > >(tee "export_tables.stdout") 2> >(tee "export_tables.stderr")
+    """
 
     stub:
-        """
-        touch stub.stdout stub.stderr stub.tsv
-        """
+    """
+    touch stub.stdout stub.stderr stub.tsv
+    """
 }
 
 process RENDER_QC_REPORT {
@@ -163,15 +128,15 @@ process RENDER_QC_REPORT {
         path("*.stderr"), emit: stderr
 
     script:
-        """
-        quarto render qc_report.qmd --to '${report_format}' \
-            > >(tee "render_${report_format}_report.stdout") 2> >(tee "render_${report_format}_report.stderr")
-        """
+    """
+    quarto render qc_report.qmd --to '${report_format}' \
+        > >(tee "render_${report_format}_report.stdout") 2> >(tee "render_${report_format}_report.stderr")
+    """
 
     stub:
-        """
-        touch "qc_report.${report_format}"
-        touch stub.stdout stub.stderr
-        """
+    """
+    touch "qc_report.${report_format}"
+    touch stub.stdout stub.stderr
+    """
 }
 
diff --git a/modules/skyline.nf b/modules/skyline.nf
@@ -138,8 +138,8 @@ process ANNOTATION_TSV_TO_CSV {
         path replicate_metadata
 
     output:
-        path("sky_annotations.csv"), emit: annotation_csv
-        path("sky_annotation_definitions.bat"), emit: annotation_definitions
+        path("metadata.annotations.csv"), emit: annotation_csv
+        path("metadata.definitions.bat"), emit: annotation_definitions
 
     shell:
     """
@@ -148,7 +148,7 @@ process ANNOTATION_TSV_TO_CSV {
 
     stub:
     """
-    touch sky_annotation_definitions.bat  sky_annotations.csv
+    touch metadata.definitions.bat metadata.annotations.csv
     """
 }
 

diff --git a/workflows/generate_qc_report.nf b/workflows/generate_qc_report.nf
@@ -1,8 +1,6 @@
 
 include { SKYLINE_RUN_REPORTS } from "../modules/skyline.nf"
 include { PARSE_REPORTS } from "../modules/qc_report.nf"
-include { NORMALIZE_DB } from "../modules/qc_report.nf"
-include { GENERATE_QC_QMD } from "../modules/qc_report.nf"
 include { RENDER_QC_REPORT } from "../modules/qc_report.nf"
 include { EXPORT_TABLES } from "../modules/qc_report.nf"
 
@@ -32,15 +30,8 @@ workflow generate_dia_qc_report {
                       precursor_report,
                       replicate_metadata)
 
-        if(params.qc_report.normalization_method != null) {
-            NORMALIZE_DB(PARSE_REPORTS.out.qc_report_db)
-            qc_report_db = NORMALIZE_DB.out.qc_report_db
-        } else {
-            qc_report_db = PARSE_REPORTS.out.qc_report_db
-        }
-
-        GENERATE_QC_QMD(qc_report_db)
-        qc_report_qmd = GENERATE_QC_QMD.out.qc_report_qmd
+        qc_report_db = PARSE_REPORTS.out.qc_report_db
+        qc_report_qmd = PARSE_REPORTS.out.qc_report_qmd
 
         report_formats = Channel.fromList(['html', 'pdf'])
         RENDER_QC_REPORT(qc_report_qmd.collect(), qc_report_db.collect(),
-Original file line number
+Diff line change
@@ Expand Up / @@ -28,7 +28,7 @@ process ENCYCLOPEDIA_SEARCH_FILE { @@
             path("${mzml_file}.features.txt"), emit: features
             path("${mzml_file}.encyclopedia.txt"), emit: results_targets
             path("${mzml_file}.encyclopedia.decoy.txt"), emit: results_decoys
         script:
         """
@@ Expand Down @@