nf-core · FriederikeHanssen · Mar 9, 2026 · Mar 9, 2026 · Mar 12, 2026 · Mar 12, 2026
@@ -18,6 +18,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- [#2154](https://github.com/nf-core/sarek/pull/2154) - Fix `--save_output_as_bam` causing duplicate emission errors, silently skipping variant calling, and running unnecessary CRAM-to-BAM conversions
 - [#2117](https://github.com/nf-core/sarek/pull/2117) - Silent failure with multi-lane samples
 - [#2143](https://github.com/nf-core/sarek/pull/2143) - Varlociraptor collecting multiple scenario files for one sample
 - [#2146](https://github.com/nf-core/sarek/pull/2146) - Fail early when `--no_intervals` is used with joint germline HaplotypeCaller

@@ -15,26 +15,8 @@
 
 process {
 
-    withName: 'CRAM_TO_BAM' {
-        ext.args   = { '-b' }
-    }
-
-    withName: 'BAM_TO_CRAM' {
-        // BAM provided for step Markduplicates either run through MD or Convert -> then saved as sorted.cram (convert) or md.cram (md directly)
-        // BAM files provided for step prepare_recal are converted and run through BQSR -> then saved as md.cram
-        // BAM files provided for step recal are converted and run through BQSR II -> then saved as md.cram
-        ext.args   = { '-C' }
-        ext.prefix = { "${meta.id}.converted" }
-        publishDir = [
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/preprocessing/converted/${meta.id}" },
-            pattern: "*{cram,crai}",
-            saveAs: { !params.save_output_as_bam ? it : null }
-        ]
-    }
-
     withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_MARKDUPLICATES|BAM_MARKDUPLICATES_SPARK):CRAM_QC_MOSDEPTH_SAMTOOLS:SAMTOOLS_STATS' {
-        ext.prefix = { "${meta.id}.md.cram" }
+        ext.prefix = { params.save_output_as_bam ? "${meta.id}.md.bam" : "${meta.id}.md.cram" }
         ext.when   = { !(params.skip_tools && params.skip_tools.split(',').contains('samtools')) }
         publishDir = [
             mode: params.publish_dir_mode,
@@ -45,9 +27,10 @@ process {
 
     withName: 'BAM_TO_CRAM_MAPPING' {
         ext.prefix = { "${meta.id}.sorted" }
-        // Run only when mapping should be saved as CRAM or when no MD is done
+        // Run only when mapping should be saved as CRAM or when no MD is done (and not saving as BAM)
         ext.when   = (params.save_mapped && !params.save_output_as_bam) ||
             (
+                !params.save_output_as_bam &&
                 (params.skip_tools && params.skip_tools.split(',').contains('markduplicates')) &&
                 !(params.tools && params.tools.split(',').contains('sentieon_dedup'))
             )
@@ -78,7 +61,7 @@ process {
             params.umi_tag ? "--BARCODE_TAG ${params.umi_tag}" :
                 (params.umi_in_read_header || params.umi_length ? "--BARCODE_TAG RX" : "")
         ].join(" ").trim() }
-        ext.prefix = { "${meta.id}.md.cram" }
+        ext.prefix = { params.save_output_as_bam ? "${meta.id}.md.bam" : "${meta.id}.md.cram" }
         ext.when   = { !(params.skip_tools && params.skip_tools.split(',').contains('markduplicates')) }
         publishDir = [
             [
@@ -87,6 +70,12 @@ process {
                 pattern: "*{cram,crai}",
                 saveAs: { !params.save_output_as_bam ? it : null }
             ],
+            [
+                mode: params.publish_dir_mode,
+                path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
+                pattern: "*{md.bam,md.bam.bai}",
+                saveAs: { params.save_output_as_bam ? it : null }
+            ],
             [
                 mode: params.publish_dir_mode,
                 path: { "${params.outdir}/reports/" },
@@ -99,42 +88,38 @@ process {
     withName: 'GATK4SPARK_MARKDUPLICATES' {
         containerOptions = ''
         ext.args   = { '--remove-sequencing-duplicates false -VS LENIENT' }
-        ext.prefix = { "${meta.id}.md.cram" }
+        ext.prefix = { params.save_output_as_bam ? "${meta.id}.md.bam" : "${meta.id}.md.cram" }
         publishDir = [
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
-            pattern: "*{cram,crai}",
-            saveAs: { !params.save_output_as_bam ? it : null }
+            [
+                mode: params.publish_dir_mode,
+                path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
+                pattern: "*{cram,crai}",
+                saveAs: { !params.save_output_as_bam ? it : null }
+            ],
+            [
+                mode: params.publish_dir_mode,
+                path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
+                pattern: "*{md.bam,md.bam.bai}",
+                saveAs: { params.save_output_as_bam ? it : null }
+            ]
         ]
     }
 
     withName: 'INDEX_MARKDUPLICATES' {
         publishDir = [
-            mode: params.publish_dir_mode,
-            path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
-            pattern: "*{cram,crai}",
-            saveAs: { !params.save_output_as_bam ? it : null }
-        ]
-    }
-
-    withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:CRAM_TO_BAM' {
-        ext.when   = { params.save_output_as_bam }
-        if (params.tools && params.tools.split(',').contains('sentieon_dedup')) {
-            ext.prefix = { "${meta.id}.dedup" }
-            publishDir = [
+            [
                 mode: params.publish_dir_mode,
-                path: { "${params.outdir}/preprocessing/sentieon_dedup/${meta.id}/" },
-                pattern: "*{dedup.bam,dedup.bam.bai}",
-                saveAs: { params.save_output_as_bam ? it : null }
-            ]
-        } else {
-            ext.prefix = { "${meta.id}.md" }
-            publishDir = [
+                path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
+                pattern: "*{cram,crai}",
+                saveAs: { !params.save_output_as_bam ? it : null }
+            ],
+            [
                 mode: params.publish_dir_mode,
                 path: { "${params.outdir}/preprocessing/markduplicates/${meta.id}/" },
-                pattern: "*{md.bam,md.bam.bai}",
+                pattern: "*{bam,bai}",
                 saveAs: { params.save_output_as_bam ? it : null }
             ]
-        }
+        ]
     }
+
 }
@@ -43,31 +43,38 @@ process {
         ]]
     }
 
-    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:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):CRAM_MERGE_INDEX_SAMTOOLS:MERGE_CRAM' {
-            ext.prefix       = { "${meta.id}.recal" }
-            ext.when         = { meta.num_intervals > 1 }
-            publishDir       = [
-                mode: params.publish_dir_mode,
-                path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },
-                pattern: "*cram",
-                saveAs: { !params.save_output_as_bam ? it : null }
+    withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):CRAM_MERGE_INDEX_SAMTOOLS:MERGE_CRAM' {
+        ext.prefix       = { "${meta.id}.recal" }
+        ext.when         = { meta.num_intervals > 1 }
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },
+            pattern: "*cram",
+            saveAs: { !params.save_output_as_bam ? it : null }
         ]
-        }
+    }
 
-        withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):CRAM_MERGE_INDEX_SAMTOOLS:INDEX_CRAM' {
-            publishDir       = [
-                mode: params.publish_dir_mode,
-                path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },
-                pattern: "*{recal.cram,recal.cram.crai}",
-                saveAs: { !params.save_output_as_bam ? it : null }
+    withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):CRAM_MERGE_INDEX_SAMTOOLS:INDEX_CRAM' {
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },
+            pattern: "*{recal.cram,recal.cram.crai}",
+            saveAs: { !params.save_output_as_bam ? it : null }
         ]
-        }
     }
 
-    withName: 'CRAM_TO_BAM_RECAL' {
+    withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):BAM_MERGE_INDEX_SAMTOOLS:MERGE_BAM' {
         ext.prefix       = { "${meta.id}.recal" }
-        ext.when         = { params.save_output_as_bam}
+        ext.when         = { meta.num_intervals > 1 }
+        publishDir       = [
+            mode: params.publish_dir_mode,
+            path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },
+            pattern: "*bam",
+            saveAs: { params.save_output_as_bam ? it : null }
+        ]
+    }
+
+    withName: 'NFCORE_SAREK:SAREK:FASTQ_PREPROCESS_GATK:(BAM_APPLYBQSR|BAM_APPLYBQSR_SPARK):BAM_MERGE_INDEX_SAMTOOLS:INDEX_MERGE_BAM' {
         publishDir       = [
             mode: params.publish_dir_mode,
             path: { "${params.outdir}/preprocessing/recalibrated/${meta.id}/" },

@@ -575,7 +575,6 @@ Expected run output:
 [88/3af664] process > NFCORE_SAREK:SAREK:BAM_APPLYBQSR:CRAM_MERGE_INDEX_SAMTOOLS:INDEX_CRAM (test)                                  [100%] 1 of 1 ✔
 [f4/828fde] process > NFCORE_SAREK:SAREK:CRAM_QC_RECAL:SAMTOOLS_STATS (test)                                                        [100%] 1 of 1 ✔
 [fb/a9d66f] process > NFCORE_SAREK:SAREK:CRAM_QC_RECAL:MOSDEPTH (test)                                                              [100%] 1 of 1 ✔
-[-        ] process > NFCORE_SAREK:SAREK:CRAM_TO_BAM_RECAL                                                                          -
 [ef/026185] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_SINGLE_STRELKA:STRELKA_SINGLE (test)  [100%] 1 of 1 ✔
 [-        ] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_SINGLE_STRELKA:MERGE_STRELKA          -
 [-        ] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_SINGLE_STRELKA:MERGE_STRELKA_GENOME   -

@@ -18,8 +18,6 @@ workflow BAM_APPLYBQSR {
 
     main:
     versions = channel.empty()
-    bam_applybqsr_single = channel.empty()
-    bam_to_merge = channel.empty()
 
     // Combine cram and intervals for spread and gather strategy
     // Move num_intervals to meta map
@@ -35,54 +33,33 @@ workflow BAM_APPLYBQSR {
         dict.map { _meta, dict_ -> [dict_] },
     )
 
-    // FOR BAMs
-    if (params.save_output_as_bam) {
+    // BAM path — populated when ext.suffix='bam', empty otherwise
+    bam_to_merge = GATK4_APPLYBQSR.out.bam
+        .map { meta, bam_ -> [groupKey(meta, meta.num_intervals), bam_] }
+        .groupTuple()
 
-        bam_applybqsr_out = GATK4_APPLYBQSR.out.bam
-            .join(GATK4_APPLYBQSR.out.bai, failOnDuplicate: true, failOnMismatch: true)
-            .branch { files ->
-                single: files[0].num_intervals == 1
-                multiple: files[0].num_intervals > 1
-            }
-
-        bam_applybqsr_single = bam_applybqsr_out.single
-
-        // For multiple intervals, gather and merge the recalibrated cram files
-        bam_to_merge = bam_applybqsr_out.multiple
-            .map { meta, bam_, _bai -> [groupKey(meta, meta.num_intervals), bam_] }
-            .groupTuple()
-    }
-
-    // Merge and index the recalibrated cram files
     BAM_MERGE_INDEX_SAMTOOLS(bam_to_merge)
 
-    // Combine single and merged multiple bam and index files, removing num_intervals field
-    bam_recal = bam_applybqsr_single
-        .mix(BAM_MERGE_INDEX_SAMTOOLS.out.bam_bai)
-        .map { meta, bam, bai -> [meta - meta.subMap('num_intervals'), bam, bai] }
-
-    // FOR CRAMs
-
-    // Gather the recalibrated cram files
+    // CRAM path — populated when ext.suffix='cram', empty otherwise
     cram_to_merge = GATK4_APPLYBQSR.out.cram.map { meta, cram_ -> [groupKey(meta, meta.num_intervals), cram_] }.groupTuple()
 
-    // Merge and index the recalibrated cram files
     CRAM_MERGE_INDEX_SAMTOOLS(
         cram_to_merge,
         fasta,
         fasta_fai,
     )
 
-    // Remove no longer necessary field: num_intervals
-    cram_recal = CRAM_MERGE_INDEX_SAMTOOLS.out.cram_crai.map { meta, cram_, crai -> [meta - meta.subMap('num_intervals'), cram_, crai] }
+    // Mix — one is always empty
+    recal_out = BAM_MERGE_INDEX_SAMTOOLS.out.bam_bai
+        .mix(CRAM_MERGE_INDEX_SAMTOOLS.out.cram_crai)
+        .map { meta, file_, index -> [meta - meta.subMap('num_intervals'), file_, index] }
 
     // Gather versions of all tools used
     versions = versions.mix(BAM_MERGE_INDEX_SAMTOOLS.out.versions)
     versions = versions.mix(CRAM_MERGE_INDEX_SAMTOOLS.out.versions)
     versions = versions.mix(GATK4_APPLYBQSR.out.versions)
 
     emit:
-    bam      = bam_recal // channel: [ meta, bam, bai ]
-    cram     = cram_recal // channel: [ meta, cram, crai ]
-    versions // channel: [ versions.yml ]
+    alignment = recal_out // channel: [ meta, file, index ] — BAM or CRAM
+    versions              // channel: [ versions.yml ]
 }
@@ -18,8 +18,6 @@ workflow BAM_APPLYBQSR_SPARK {
 
     main:
     versions = channel.empty()
-    bam_applybqsr_single = channel.empty()
-    bam_to_merge = channel.empty()
 
     // Combine cram and intervals for spread and gather strategy
     // Move num_intervals to meta map
@@ -35,54 +33,33 @@ workflow BAM_APPLYBQSR_SPARK {
         dict.map { _meta, dict_ -> [dict_] },
     )
 
-    // FOR BAMs
-    if (params.save_output_as_bam) {
+    // BAM path — populated when ext.suffix='bam', empty otherwise
+    bam_to_merge = GATK4SPARK_APPLYBQSR.out.bam
+        .map { meta, bam_ -> [groupKey(meta, meta.num_intervals), bam_] }
+        .groupTuple()
 
-        bam_applybqsr_out = GATK4SPARK_APPLYBQSR.out.bam
-            .join(GATK4SPARK_APPLYBQSR.out.bai, failOnDuplicate: true, failOnMismatch: true)
-            .branch { files ->
-                single: files[0].num_intervals == 1
-                multiple: files[0].num_intervals > 1
-            }
-
-        bam_applybqsr_single = bam_applybqsr_out.single
-
-        // For multiple intervals, gather and merge the recalibrated cram files
-        bam_to_merge = bam_applybqsr_out.multiple
-            .map { meta, bam_, _bai -> [groupKey(meta, meta.num_intervals), bam_] }
-            .groupTuple()
-    }
-
-    // Merge and index the recalibrated cram files
     BAM_MERGE_INDEX_SAMTOOLS(bam_to_merge)
 
-    // Combine single and merged multiple bam and index files, removing num_intervals field
-    bam_recal = bam_applybqsr_single
-        .mix(BAM_MERGE_INDEX_SAMTOOLS.out.bam_bai)
-        .map { meta, bam, bai -> [meta - meta.subMap('num_intervals'), bam, bai] }
-
-    // FOR CRAMs
-
-    // Gather the recalibrated cram files
+    // CRAM path — populated when ext.suffix='cram', empty otherwise
     cram_to_merge = GATK4SPARK_APPLYBQSR.out.cram.map { meta, cram_ -> [groupKey(meta, meta.num_intervals), cram_] }.groupTuple()
 
-    // Merge and index the recalibrated cram files
     CRAM_MERGE_INDEX_SAMTOOLS(
         cram_to_merge,
         fasta,
         fasta_fai,
     )
 
-    // Remove no longer necessary field: num_intervals
-    cram_recal = CRAM_MERGE_INDEX_SAMTOOLS.out.cram_crai.map { meta, cram_, crai -> [meta - meta.subMap('num_intervals'), cram_, crai] }
+    // Mix — one is always empty
+    recal_out = BAM_MERGE_INDEX_SAMTOOLS.out.bam_bai
+        .mix(CRAM_MERGE_INDEX_SAMTOOLS.out.cram_crai)
+        .map { meta, file_, index -> [meta - meta.subMap('num_intervals'), file_, index] }
 
     // Gather versions of all tools used
     versions = versions.mix(BAM_MERGE_INDEX_SAMTOOLS.out.versions)
     versions = versions.mix(CRAM_MERGE_INDEX_SAMTOOLS.out.versions)
     versions = versions.mix(GATK4SPARK_APPLYBQSR.out.versions)
 
     emit:
-    bam      = bam_recal // channel: [ meta, bam, bai ]
-    cram     = cram_recal // channel: [ meta, cram, crai ]
-    versions // channel: [ versions.yml ]
+    alignment = recal_out // channel: [ meta, file, index ] — BAM or CRAM
+    versions              // channel: [ versions.yml ]
 }
@@ -4,8 +4,9 @@
 // For all modules here:
 // A when clause condition is defined in the conf/modules.config to determine if the module should be run
 
-include { CRAM_QC_MOSDEPTH_SAMTOOLS } from '../cram_qc_mosdepth_samtools/main'
-include { GATK4_MARKDUPLICATES      } from '../../../modules/nf-core/gatk4/markduplicates/main'
+include { CRAM_QC_MOSDEPTH_SAMTOOLS                               } from '../cram_qc_mosdepth_samtools/main'
+include { GATK4_MARKDUPLICATES                                    } from '../../../modules/nf-core/gatk4/markduplicates/main'
+include { SAMTOOLS_INDEX                  as INDEX_MARKDUPLICATES } from '../../../modules/nf-core/samtools/index/main'
 
 workflow BAM_MARKDUPLICATES {
     take:
@@ -21,22 +22,29 @@ workflow BAM_MARKDUPLICATES {
     // RUN MARKUPDUPLICATES
     GATK4_MARKDUPLICATES(bam, fasta.map{ meta, fasta -> [ fasta ] }, fasta_fai.map{ meta, fasta_fai -> [ fasta_fai ] })
 
-    // Join with the crai file
-    cram = GATK4_MARKDUPLICATES.out.cram.join(GATK4_MARKDUPLICATES.out.crai, failOnDuplicate: true, failOnMismatch: true)
+    // BAM path: module does not auto-index BAM output, so index explicitly
+    INDEX_MARKDUPLICATES(GATK4_MARKDUPLICATES.out.bam)
 
-    // QC on CRAM
-    CRAM_QC_MOSDEPTH_SAMTOOLS(cram, fasta, intervals_bed_combined)
+    // Unified alignment output — BAM or CRAM depending on save_output_as_bam
+    alignment = GATK4_MARKDUPLICATES.out.bam
+        .join(INDEX_MARKDUPLICATES.out.bai, failOnDuplicate: true, failOnMismatch: true)
+        .mix(GATK4_MARKDUPLICATES.out.cram
+            .join(GATK4_MARKDUPLICATES.out.crai, failOnDuplicate: true, failOnMismatch: true))
+
+    // QC on alignment
+    CRAM_QC_MOSDEPTH_SAMTOOLS(alignment, fasta, intervals_bed_combined)
 
     // Gather all reports generated
     reports = reports.mix(GATK4_MARKDUPLICATES.out.metrics)
     reports = reports.mix(CRAM_QC_MOSDEPTH_SAMTOOLS.out.reports)
 
     // Gather versions of all tools used
     versions = versions.mix(GATK4_MARKDUPLICATES.out.versions)
+    versions = versions.mix(INDEX_MARKDUPLICATES.out.versions)
     versions = versions.mix(CRAM_QC_MOSDEPTH_SAMTOOLS.out.versions)
 
     emit:
-    cram
+    alignment   // channel: [ meta, file, index ] — BAM or CRAM
     reports
 
     versions    // channel: [ versions.yml ]