Merge remote-tracking branch 'origin/dev' into multiqc_channels

nf-core · Jul 20, 2022 · d83794f · d83794f
2 parents cbe9755 + e5a4ffb
commit d83794f
Show file tree

Hide file tree

Showing 35 changed files with 1,254 additions and 693 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -43,6 +43,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#632](https://github.com/nf-core/sarek/pull/632) - Added params `--snpeff_version` to allow more configuration on the snpeff container definition
 - [#632](https://github.com/nf-core/sarek/pull/632) - Added params `--vep_include_fasta` to use the fasta file for annotation
 - [#639](https://github.com/nf-core/sarek/pull/639) - Adding genes-txt-file and summary-html-file to the published output from snpEff.
+- [#647](https://github.com/nf-core/sarek/pull/647) - Update resource requests for preprocessing based on what worked for 5 ICGC matched WGS samples
 - [#652](https://github.com/nf-core/sarek/pull/652) - Added full size somatic test profile.
 
 ### Changed
@@ -89,11 +90,17 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#630](https://github.com/nf-core/sarek/pull/630) - Update citations file
 - [#632](https://github.com/nf-core/sarek/pull/632) - Update `snpEff` version to `5.1` and cache up to `105`
 - [#632](https://github.com/nf-core/sarek/pull/632) - Update `VEP` version to `106.1` and cache up to `106`
+- [#618](https://github.com/nf-core/sarek/pull/618) - Update `multiqc` module update test yml files
+- [#618](https://github.com/nf-core/sarek/pull/618) - Update test yml files
 - [#633](https://github.com/nf-core/sarek/pull/633) - Update `BCFTOOLS` version to `1.15.1`
 - [#644](https://github.com/nf-core/sarek/pull/644) - Use `-Y` for `bwa-mem(2)` and remove `-M`
 - [#645](https://github.com/nf-core/sarek/pull/645) - Merge `tests/nextflow.config` in `conf/test.config`
 - [#646](https://github.com/nf-core/sarek/pull/646) - Update `nextflow_schema.json` to reflect new parameters and functions, removes `--annotation_cache`, removes `--ascat_chromosomes`
+- [#649](https://github.com/nf-core/sarek/pull/649) - Update, simplify and add more files to all `test_*.yml` files
+- [#651](https://github.com/nf-core/sarek/pull/651) - Added TIDDIT_SOMATIC subworkflow
 - [#653](https://github.com/nf-core/sarek/pull/653) - Coherent results subfolder structure between preprocessing, variantcalling and reporting
+- [#659](https://github.com/nf-core/sarek/pull/659) - Update usage.md docu section on `How to run ASCAT with WES`
+- [#663](https://github.com/nf-core/sarek/pull/663) - Add separate parameters for `ASCAT` and `ControlFREEC` back in
 
 ### Fixed
 
@@ -143,6 +150,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#642](https://github.com/nf-core/sarek/pull/642) - Only unzip ref files if tool is run, only publish ref files if `--save_reference` and simplify CNKit logic
 - [#650](https://github.com/nf-core/sarek/pull/650) - Fix intervals checks
 - [#654](https://github.com/nf-core/sarek/pull/654) - Allow any step but annotation to start from BAM files
+- [#658](https://github.com/nf-core/sarek/pull/658) - Fix split fastq names in multiqc-report
 - [#666](https://github.com/nf-core/sarek/pull/666) - Simplify multiqc config channel input
 
 ### Deprecated

diff --git a/assets/multiqc_config.yml b/assets/multiqc_config.yml
@@ -59,6 +59,10 @@ extra_fn_clean_exts:
     pattern: "^.*.(md|recal).mosdepth.(global|region).dist"
     module: mosdepth
 
+sample_names_replace_regex: True
+sample_names_replace:
+  ".[0-9]{4}": ".md"
+
 sp:
   snpeff:
     contents: "SnpEff_version"

diff --git a/conf/base.config b/conf/base.config
@@ -49,8 +49,28 @@ process {
         errorStrategy = 'retry'
         maxRetries    = 2
     }
-    withName: 'BWAMEM2_MEM' {
-        memory       = { check_max( 60.GB * task.attempt, 'memory' ) }
+    withName: 'BWAMEM1_MEM|BWAMEM2_MEM' {
+        cpus         = { check_max( 24 * task.attempt, 'cpus' ) }
+        memory       = { check_max( 36.GB * task.attempt, 'memory' ) }
         time         = { check_max( 48.h  * task.attempt, 'time'    ) }
     }
+    withName: 'FASTP'{
+        cpus         = { check_max( 12 * task.attempt, 'cpus' ) }
+    }
+    withName:'FASTQC|FASTP|MOSDEPTH|SAMTOOLS_CONVERT|SAMTOOLS_MERGE'{
+        memory       = { check_max( 4.GB * task.attempt, 'memory' ) }
+    }
+    withName:'GATK4_APPLYBQSR|GATK4_APPLYBQSR_SPARK|GATK4_BASERECALIBRATOR|SAMTOOLS_STATS'{
+        cpus         = { check_max( 4 * task.attempt, 'cpus' ) }
+    }
+    withName:'GATK4_APPLYBQSR|GATK4_APPLYBQSR_SPARK|GATK4_BASERECALIBRATOR|GATK4_GATHERBQSRREPORTS'{
+        memory       = { check_max( 46.GB * task.attempt, 'memory' ) }
+    }
+    withName: 'GATK4_MARKDUPLICATES'{
+        memory       = { check_max( 300.GB * task.attempt, 'memory' ) }
+    }
+    withName: 'FREEBAYES|SAMTOOLS_STATS|SAMTOOLS_INDEX|UNZIP' {
+        cpus         = { check_max( 1 * task.attempt, 'cpus'    ) }
+    }
+
 }
diff --git a/conf/modules.config b/conf/modules.config
@@ -181,7 +181,7 @@ process {
     }
 
     withName: 'TABIX_BGZIPTABIX_INTERVAL_SPLIT' {
-        ext.prefix       = {"${meta.id}.bed"}
+        ext.prefix       = {"${meta.id}"}
         publishDir       = [
             enabled: params.save_reference,
             mode: params.publish_dir_mode,
@@ -446,7 +446,7 @@ process {
     }
 
 
-    if (params.step == 'mapping' || params.step == 'markduplicates'|| params.step == 'prepare_recalibration'|| params.step == 'recalibrate') {
+    if ((params.step == 'mapping' || params.step == 'markduplicates'|| params.step == 'prepare_recalibration'|| params.step == 'recalibrate') && (!(params.skip_tools && params.skip_tools.split(',').contains('baserecalibrator')))) {
         withName: 'NFCORE_SAREK:SAREK:(RECALIBRATE|RECALIBRATE_SPARK):MERGE_INDEX_CRAM:MERGE_CRAM' {
             ext.prefix       = { "${meta.id}.recal" }
             ext.when         = { meta.num_intervals > 1 }
@@ -529,7 +529,7 @@ process {
         ]
     }
 
-    if (params.step == 'mapping' || params.step == 'markduplicates'|| params.step == 'prepare_recalibration'|| params.step == 'recalibrate') {
+    if ((params.step == 'mapping' || params.step == 'markduplicates'|| params.step == 'prepare_recalibration'|| params.step == 'recalibrate') && (!(params.skip_tools && params.skip_tools.split(',').contains('baserecalibrator')))) {
         withName: 'NFCORE_SAREK:SAREK:CRAM_QC:MOSDEPTH' {
             ext.prefix       = { "${meta.id}.recal" }
         }
@@ -794,7 +794,7 @@ process{
                         forcegccontentnormalization: params.wes ? "1" : "0",
                         minimalsubclonepresence: params.wes ? "30" : "20",
                         noisydata: params.wes ? "TRUE" : "FALSE",
-                        ploidy: params.ploidy,
+                        ploidy: params.cf_ploidy,
                         printNA:  params.wes ? "FALSE" : "TRUE",
                         readcountthreshold:  params.wes ? "50" : "10",
                         sex: meta.sex,
@@ -811,7 +811,7 @@ process{
     }
 
     withName: 'FREEC2BED' {
-        ext.args = { "${params.ploidy}" }
+        ext.args = { "${params.cf_ploidy}" }
         publishDir       = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/controlfreec/${meta.id}/" },
@@ -820,7 +820,7 @@ process{
     }
 
     withName: 'FREEC2CIRCOS' {
-        ext.args = { "${params.ploidy}" }
+        ext.args = { "${params.cf_ploidy}" }
         publishDir       = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/controlfreec/${meta.id}/" },
@@ -829,7 +829,7 @@ process{
     }
 
     withName: 'MAKEGRAPH' {
-        ext.args = { "${params.ploidy}" }
+        ext.args = { "${params.cf_ploidy}" }
         publishDir       = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/variant_calling/controlfreec/${meta.id}/" },
@@ -958,7 +958,7 @@ process{
             "gender": meta.sex,
             "genomeVersion": params.ascat_genome,
             "purity": params.ascat_purity,
-            "ploidy": params.ploidy,
+            "ploidy": params.ascat_ploidy,
             "minCounts": params.ascat_min_counts,
             "chrom_names": meta.sex == 'XX' ? "c(1:22, 'X')" : "c(1:22, 'X', 'Y')",
             "min_base_qual": params.ascat_min_base_qual,
@@ -1028,7 +1028,7 @@ process{
                         forcegccontentnormalization: params.wes ? "1" : "0",
                         minimalsubclonepresence: params.wes ? "30" : "20",
                         noisydata: params.wes ? "TRUE" : "FALSE",
-                        ploidy: params.ploidy,
+                        ploidy: params.cf_ploidy,
                         printNA:  params.wes ? "FALSE" : "TRUE",
                         readcountthreshold:  params.wes ? "50" : "10",
                         sex: meta.sex,
@@ -1091,21 +1091,21 @@ process{
 
     if (params.tools && params.tools.split(',').contains('tiddit')) {
     //TIDDIT
-        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:RUN_TIDDIT_NORMAL:TABIX_BGZIP_TIDDIT_SV' {
+        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:RUN_TIDDIT_SOMATIC:RUN_TIDDIT_NORMAL:TABIX_BGZIP_TIDDIT_SV' {
             ext.prefix       = {"${meta.id}.tiddit.normal"}
         }
 
-        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:RUN_TIDDIT_TUMOR:TABIX_BGZIP_TIDDIT_SV' {
+        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:RUN_TIDDIT_SOMATIC:RUN_TIDDIT_TUMOR:TABIX_BGZIP_TIDDIT_SV' {
             ext.prefix       = {"${meta.id}.tiddit.tumor"}
         }
 
         //SVDB
-        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:SVDB_MERGE' {
+        withName: 'NFCORE_SAREK:SAREK:PAIR_VARIANT_CALLING:RUN_TIDDIT_SOMATIC:SVDB_MERGE' {
             ext.prefix       = { "${meta.id}.tiddit" }
             publishDir       = [
                 mode: params.publish_dir_mode,
                 path: { "${params.outdir}/variant_calling/tiddit/${meta.id}/" },
-                pattern: "*vcf"
+                pattern: "*{vcf,vcf.gz}"
             ]
         }
     }

diff --git a/conf/test.config b/conf/test.config
@@ -16,7 +16,6 @@ try {
 }
 
 params {
-
     config_profile_name        = 'Test profile'
     config_profile_description = 'Minimal test dataset to check pipeline function'