From d4422a10a166ddb67e302b56169e4dbb781b6c1b Mon Sep 17 00:00:00 2001
From: Shihab Dider <shihabdider@gmail.com>
Date: Thu, 17 Oct 2024 13:58:21 -0400
Subject: [PATCH] feat: wire up estimatelibrarycomplexity to bam_qc

---
 conf/modules.config                           |  8 ++++++++
 .../local/{bam_qc_picard => bam_qc}/main.nf   | 20 +++++++++++++++----
 tests/test_runs/chr21_test/params.json        |  2 +-
 workflows/nfcasereports.nf                    | 11 ++++++----
 4 files changed, 32 insertions(+), 9 deletions(-)
 rename subworkflows/local/{bam_qc_picard => bam_qc}/main.nf (63%)

diff --git a/conf/modules.config b/conf/modules.config
index 0325c71..6900ab0 100644
--- a/conf/modules.config
+++ b/conf/modules.config
@@ -53,6 +53,14 @@ process {
         ]
     }
 
+    withName: 'GATK4_ESTIMATELIBRARYCOMPLEXITY' {
+        publishDir = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/qc_reports/gatk/${meta.id}" },
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename },
+        ]
+    }
+
     withName: 'SAMTOOLS_STATS' {
         ext.when         = { !(params.skip_tools && params.skip_tools.split(',').contains('samtools')) }
         ext.prefix       = { "${meta.id}.sorted.cram" }
diff --git a/subworkflows/local/bam_qc_picard/main.nf b/subworkflows/local/bam_qc/main.nf
similarity index 63%
rename from subworkflows/local/bam_qc_picard/main.nf
rename to subworkflows/local/bam_qc/main.nf
index 8ae0e7e..b0333e4 100644
--- a/subworkflows/local/bam_qc_picard/main.nf
+++ b/subworkflows/local/bam_qc/main.nf
@@ -4,14 +4,16 @@
 
 include { PICARD_COLLECTWGSMETRICS     } from '../../../modules/nf-core/picard/collectwgsmetrics/main'
 include { PICARD_COLLECTMULTIPLEMETRICS           } from '../../../modules/nf-core/picard/collectmultiplemetrics/main'
+include { GATK4_ESTIMATELIBRARYCOMPLEXITY           } from '../../../modules/nf-core/gatk4/estimatelibrarycomplexity/main'
 
-fasta                               = WorkflowNfcasereports.create_file_channel(params.fasta)
-fai                           = WorkflowNfcasereports.create_file_channel(params.fasta_fai)
-intervals                          = WorkflowNfcasereports.create_file_channel(params.intervals)
+fasta = WorkflowNfcasereports.create_file_channel(params.fasta)
+fai = WorkflowNfcasereports.create_file_channel(params.fasta_fai)
+intervals = WorkflowNfcasereports.create_file_channel(params.intervals)
 
-workflow BAM_QC_PICARD {
+workflow BAM_QC {
     take:
     bam                         // channel: [mandatory] [ meta, bam, bai ]
+    dict
 
     main:
     versions = Channel.empty()
@@ -30,13 +32,23 @@ workflow BAM_QC_PICARD {
         fai.map{ it -> [ [ id:'fai' ], it ] },
     )
 
+    bam_only = bam.map{ meta, bam, bai -> [ meta, bam ] }
+    GATK4_ESTIMATELIBRARYCOMPLEXITY(
+        bam_only,
+        fasta,
+        fai,
+        dict
+    )
+
     // Gather all reports generated
     reports = reports.mix(PICARD_COLLECTWGSMETRICS.out.metrics)
     reports = reports.mix(PICARD_COLLECTMULTIPLEMETRICS.out.metrics)
+    reports = reports.mix(GATK4_ESTIMATELIBRARYCOMPLEXITY.out.metrics)
 
     // Gather versions of all tools used
     versions = versions.mix(PICARD_COLLECTWGSMETRICS.out.versions)
     versions = versions.mix(PICARD_COLLECTMULTIPLEMETRICS.out.versions)
+    versions = versions.mix(GATK4_ESTIMATELIBRARYCOMPLEXITY.out.versions)
 
     emit:
     reports
diff --git a/tests/test_runs/chr21_test/params.json b/tests/test_runs/chr21_test/params.json
index f065504..03dd3d5 100644
--- a/tests/test_runs/chr21_test/params.json
+++ b/tests/test_runs/chr21_test/params.json
@@ -4,7 +4,7 @@
     "fasta": "/gpfs/commons/home/sdider/DB/GATK/human_g1k_v37_decoy.fasta",
     "fasta_fai": "/gpfs/commons/home/sdider/DB/GATK/human_g1k_v37_decoy.fasta.fai",
     "bwa": "/gpfs/commons/home/sdider/DB/GATK/bwa/",
-    "tools": "snv_multiplicity",
+    "tools": "bamqc",
     "outdir": "./results",
     "pon_dryclean": "/gpfs/commons/home/sdider/Projects/nf-casereports/tests/test_data/chr21_pon.rds",
     "field_dryclean": "reads",
diff --git a/workflows/nfcasereports.nf b/workflows/nfcasereports.nf
index bc6352c..8aae161 100644
--- a/workflows/nfcasereports.nf
+++ b/workflows/nfcasereports.nf
@@ -498,7 +498,7 @@ include { CRAM_QC_MOSDEPTH_SAMTOOLS as CRAM_QC_NO_MD } from '../subworkflows/loc
 include { CRAM_QC_MOSDEPTH_SAMTOOLS as CRAM_QC_RECAL } from '../subworkflows/local/cram_qc_mosdepth_samtools/main'
 
 // BAM Picard QC
-include { BAM_QC_PICARD } from '../subworkflows/local/bam_qc_picard/main'
+include { BAM_QC } from '../subworkflows/local/bam_qc/main'
 
 // Create recalibration tables
 include { BAM_BASERECALIBRATOR } from '../subworkflows/local/bam_baserecalibrator/main'
@@ -876,11 +876,14 @@ workflow NFCASEREPORTS {
         bam_qc_calling = alignment_bams_final
             .join(bam_qc_inputs)
             .map { it -> [ it[1], it[2], it[3] ] } // meta, bam, bai
-        BAM_QC_PICARD(bam_qc_calling)
+
+        // omit meta since it is not used in the BAM_QC
+        dict_path = dict.map{ meta, dict -> dict }
+        BAM_QC(bam_qc_calling, dict_path)
 
         // Gather QC
-        reports = reports.mix(BAM_QC_PICARD.out.reports.collect{ meta, report -> report })
-        versions = versions.mix(BAM_QC_PICARD.out.versions)
+        reports = reports.mix(BAM_QC.out.reports.collect{ meta, report -> report })
+        versions = versions.mix(BAM_QC.out.versions)
     }
 
     // SV Calling