update spark modules

CHANGELOG.md

@@ -57,6 +57,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#571](https://github.com/nf-core/sarek/pull/571) - Including and using GATK4's mergeVcfs. Removing the local module `concat_vcf`.
 - [#572](https://github.com/nf-core/sarek/pull/572) - Adjusted subway map svg for firefox compatibility
 - [#578](https://github.com/nf-core/sarek/pull/578) - Updated module deeptools/bamcoverage
+- [#585](https://github.com/nf-core/sarek/pull/585) - Remove explicit BAM to CRAM conversion after MarkduplicatesSpark; tool does it internally
 
 ### Fixed
 
@@ -82,6 +83,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#524](https://github.com/nf-core/sarek/pull/524) - Fix intervals usage by counting the actual list of scatter/gather files produced and not overall number of intervals
 - [#549](https://github.com/nf-core/sarek/pull/549) - Fix unique lanes required for Freebayes: issue [#311](https://github.com/nf-core/sarek/issues/311), replaces `meta.clone()` with actual copy of map to avoid issues with <https://nfcore.slack.com/archives/C027CM7P08M/p1644241819942339>
 - [#567](https://github.com/nf-core/sarek/pull/567) - Fix interval name resolving during scatter/gather by moving logic to modules.config causing name to be correctly resolved on process execution; also fixed duplicate naming when variant callers produce multiple vcf files by adding field `type` to `meta` map
+- [#585](https://github.com/nf-core/sarek/pull/585) - Fix Spark usage for GATK4 modules.
 
 ### Deprecated
 

conf/modules.config

@@ -339,8 +339,7 @@ process {
 
     withName: 'GATK4_MARKDUPLICATES_SPARK' {
         ext.args         = '--remove-sequencing-duplicates false -VS LENIENT'
-        //TODO after step markduplicates we only deal with cram now, so this is not necessary, fix in PR that updates spark modules
-        ext.prefix       = { !(params.skip_tools && (params.skip_tools.contains('bamqc') || params.skip_tools.contains('deeptools'))) ? "${meta.id}.md.bam" : "${meta.id}.md.cram" }
+        ext.prefix       = { "${meta.id}.md.cram" }
         publishDir       = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/preprocessing/${meta.id}/markduplicates" },

modules.json

@@ -139,7 +139,7 @@
                 "git_sha": "df2620cfc7e4c21b14ed03c1c928f09fbabf83c4"
             },
             "gatk4/markduplicatesspark": {
-                "git_sha": "e04970b7d249365cafa5a52912f9a28840481c05"
+                "git_sha": "cda434611f851bbc4fe3d876e6b702cf721a0175"
             },
             "gatk4/mergemutectstats": {
                 "git_sha": "169b2b96c1167f89ab07127b7057c1d90a6996c7"

nextflow.config

@@ -145,7 +145,7 @@ profiles {
     }
     docker {
         docker.enabled         = true
-        docker.userEmulation   = true
+        docker.userEmulation   = false
         singularity.enabled    = false
         podman.enabled         = false
         shifter.enabled        = false