Merge pull request #1123 from nf-core/display_sentieon_dedup_metrics_…

…in_multiqc Display sentieon dedup metrics in multiqc
nf-core · Jun 21, 2023 · e3c16f6 · e3c16f6
2 parents 67499b6 + aa1ce9a
commit e3c16f6
Show file tree

Hide file tree

Showing 8 changed files with 47 additions and 3 deletions.
diff --git a/assets/multiqc_config.yml b/assets/multiqc_config.yml
@@ -66,7 +66,38 @@ sample_names_replace:
   "\\.[0-9]{4}$": ".md" # should match ".0001" but only at the end of strings for module Markduplicates/EstimateLibraryComplexity
   module: picard
 
+custom_data:
+  dedup_metrics:
+    id: "dedup_metrics"
+    section_name: "Sentieon Dedup Metrics"
+    plot_type: "table"
+    pconfig:
+      id: "dedup_metrics"
+      namespace: "Sentieon Dedup Metrics"
+    headers:
+      LIBRARY:
+        description: "LIBRARY"
+      UNPAIRED_READS_EXAMINED:
+        description: "UNPAIRED_READS_EXAMINE"
+      READ_PAIRS_EXAMINED:
+        description: "READ_PAIRS_EXAMINED"
+      SECONDARY_OR_SUPPLEMENTARY_RDS:
+        description: "SECONDARY_OR_SUPPLEMENTARY_RDS"
+      UNMAPPED_READS:
+        description: "UNMAPPED_READS"
+      UNPAIRED_READ_DUPLICATES:
+        description: "UNPAIRED_READ_DUPLICATES"
+      READ_PAIR_DUPLICATES:
+        description: "READ_PAIR_DUPLICATES"
+      READ_PAIR_OPTICAL_DUPLICATES:
+        description: "READ_PAIR_OPTICAL_DUPLICATES"
+      PERCENT_DUPLICATION:
+        description: "PERCENT_DUPLICATION"
+      ESTIMATED_LIBRARY_SIZE:
+        description: "ESTIMATED_LIBRARY_SIZE"
 sp:
   snpeff:
     contents: "SnpEff_version"
     max_filesize: 5000000
+  dedup_metrics:
+    fn: "*.metrics.multiqc.tsv"
diff --git a/conf/modules/sentieon_dedup.config b/conf/modules/sentieon_dedup.config
@@ -28,7 +28,7 @@ process {
             [
                 mode: params.publish_dir_mode,
                 path: { "${params.outdir}/reports/" },
-                pattern: "*metrics",
+                pattern: "*{metrics,metrics.multiqc.tsv}",
                 saveAs: { !(params.skip_tools && params.skip_tools.split(',').contains('sentieon_dedup_report')) ? "sentieon_dedup/${meta.id}/${it}" : null}
             ]
         ]

diff --git a/modules.json b/modules.json
@@ -383,7 +383,7 @@
                     },
                     "sentieon/dedup": {
                         "branch": "master",
-                        "git_sha": "2cb2012b268b3502ac74463e510e76be99e2cdbf",
+                        "git_sha": "e311732e4d878d7e47872226bef5f8630dc81675",
                         "installed_by": ["modules"]
                     },
                     "sentieon/gvcftyper": {

diff --git a/modules/nf-core/sentieon/dedup/main.nf b/modules/nf-core/sentieon/dedup/main.nf
diff --git a/modules/nf-core/sentieon/dedup/meta.yml b/modules/nf-core/sentieon/dedup/meta.yml
diff --git a/subworkflows/local/bam_sentieon_dedup/main.nf b/subworkflows/local/bam_sentieon_dedup/main.nf
@@ -34,6 +34,7 @@ workflow BAM_SENTIEON_DEDUP {
 
     // Gather all reports generated
     reports = reports.mix(SENTIEON_DEDUP.out.metrics)
+    reports = reports.mix(SENTIEON_DEDUP.out.metrics_multiqc_tsv)
     reports = reports.mix(SENTIEON_DEDUP.out.score)
     reports = reports.mix(CRAM_QC_MOSDEPTH_SAMTOOLS.out.reports)
 

diff --git a/tests/test_sentieon_dedup_from_bam.yml b/tests/test_sentieon_dedup_from_bam.yml
@@ -13,6 +13,8 @@
     - path: results/csv/recalibrated.csv
       md5sum: 1888a924bc70bd80165a96ad641e22d6
     - path: results/multiqc
+    - path: results/multiqc/multiqc_report.html
+      contains: ["Sentieon Dedup Metrics", "PERCENT_DUPLICATION", "ESTIMATED_LIBRARY_SIZE"]
     - path: results/preprocessing/sentieon_dedup/test/test.md.cram
     # binary changes md5sums on reruns
     - path: results/preprocessing/sentieon_dedup/test/test.md.cram.crai

diff --git a/tests/test_sentieon_dedup_from_cram.yml b/tests/test_sentieon_dedup_from_cram.yml
@@ -13,6 +13,8 @@
     - path: results/csv/recalibrated.csv
       md5sum: 1888a924bc70bd80165a96ad641e22d6
     - path: results/multiqc
+    - path: results/multiqc/multiqc_report.html
+      contains: ["Sentieon Dedup Metrics", "PERCENT_DUPLICATION", "ESTIMATED_LIBRARY_SIZE"]
     - path: results/preprocessing/sentieon_dedup/test/test.md.cram
     # binary changes md5sums on reruns
     - path: results/preprocessing/sentieon_dedup/test/test.md.cram.crai