feat(rustar): wire rustar-aligner as opt-in STAR drop-in [skip ci]

pinin4fjords · claude · pinin4fjords · commit fca6a654f9e7 · 2026-05-12T14:55:33.000+01:00
Add local RUSTAR_ALIGN and RUSTAR_GENOMEGENERATE modules using
ghcr.io/scverse/rustar-aligner. Both reuse STAR's CLI and on-disk index
format, so the dispatch in align_star and prepare_genome_indices just
gets one more conditional. The new --use_rustar_star toggle mirrors the
existing --use_sentieon_star / --use_parabricks_star pattern.

Tests, the cross-aligner comparison harness, and on-VM verification
land in follow-up commits on this branch.

Co-Authored-By: Claude Opus 4.7 (1M context) &lt;noreply@anthropic.com&gt;
diff --git a/conf/modules/align_star.config b/conf/modules/align_star.config
@@ -3,7 +3,7 @@ process {
     // STAR alignment configuration
     // Conditionals for when these processes run are handled in the workflow
 
-    withName: '.*ALIGN_STAR:STAR_ALIGN|.*ALIGN_STAR:SENTIEON_STAR_ALIGN|.*ALIGN_STAR:PARABRICKS_RNA_FQ2BAM' {
+    withName: '.*ALIGN_STAR:STAR_ALIGN|.*ALIGN_STAR:SENTIEON_STAR_ALIGN|.*ALIGN_STAR:PARABRICKS_RNA_FQ2BAM|.*ALIGN_STAR:RUSTAR_ALIGN' {
         ext.args = {
             def isPbrun = task.process.contains('PARABRICKS')
             def isSentieon = task.process.endsWith(':SENTIEON_STAR_ALIGN')
diff --git a/main.nf b/main.nf
@@ -111,6 +111,7 @@ workflow NFCORE_RNASEQ {
         params.skip_pseudo_alignment,
         params.use_sentieon_star,
         params.use_parabricks_star,
+        params.use_rustar_star,
         isStarIndexLegacy() ?: false
     )
 
diff --git a/modules/local/rustar_align/align/main.nf b/modules/local/rustar_align/align/main.nf
@@ -0,0 +1,90 @@
+process RUSTAR_ALIGN {
+    tag "$meta.id"
+    label 'process_high'
+
+    container "ghcr.io/scverse/rustar-aligner:dev"
+
+    input:
+    tuple val(meta), path(reads, stageAs: "input*/*")
+    tuple val(meta2), path(index)
+    tuple val(meta3), path(gtf)
+    val star_ignore_sjdbgtf
+
+    output:
+    tuple val(meta), path('*Log.final.out')   , emit: log_final
+    tuple val(meta), path('*Log.out')         , emit: log_out
+    tuple val(meta), path('*Log.progress.out'), emit: log_progress
+    tuple val("${task.process}"), val('rustar-aligner'), eval("rustar-aligner --version | sed -n '1{s/^rustar-aligner //;p}'"), emit: versions_rustar, topic: versions
+
+    tuple val(meta), path('*d.out.bam')                              , optional:true, emit: bam
+    tuple val(meta), path("${prefix}.sortedByCoord.out.bam")         , optional:true, emit: bam_sorted
+    tuple val(meta), path("${prefix}.Aligned.sortedByCoord.out.bam") , optional:true, emit: bam_sorted_aligned
+    tuple val(meta), path('*toTranscriptome.out.bam')                , optional:true, emit: bam_transcript
+    tuple val(meta), path('*Aligned.unsort.out.bam')                 , optional:true, emit: bam_unsorted
+    tuple val(meta), path('*fastq.gz')                               , optional:true, emit: fastq
+    tuple val(meta), path('*.tab')                                   , optional:true, emit: tab
+    tuple val(meta), path('*.SJ.out.tab')                            , optional:true, emit: spl_junc_tab
+    tuple val(meta), path('*.ReadsPerGene.out.tab')                  , optional:true, emit: read_per_gene_tab
+    tuple val(meta), path('*.out.junction')                          , optional:true, emit: junction
+    tuple val(meta), path('*.out.sam')                               , optional:true, emit: sam
+    tuple val(meta), path('*.wig')                                   , optional:true, emit: wig
+    tuple val(meta), path('*.bg')                                    , optional:true, emit: bedgraph
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}"
+    def reads1 = []
+    def reads2 = []
+    meta.single_end ? [reads].flatten().each{ read -> reads1 << read} : reads.eachWithIndex{ v, ix -> ( ix & 1 ? reads2 : reads1) << v }
+    def ignore_gtf      = star_ignore_sjdbgtf ? '' : "--sjdbGTFfile $gtf"
+    attrRG          = args.contains("--outSAMattrRGline") ? "" : "--outSAMattrRGline 'ID:$prefix' 'SM:$prefix'"
+    def out_sam_type    = (args.contains('--outSAMtype')) ? '' : '--outSAMtype BAM Unsorted'
+    mv_unsorted_bam = (args.contains('--outSAMtype BAM Unsorted SortedByCoordinate')) ? "mv ${prefix}.Aligned.out.bam ${prefix}.Aligned.unsort.out.bam" : ''
+    """
+    rustar-aligner \\
+        --genomeDir $index \\
+        --readFilesIn ${reads1.join(",")} ${reads2.join(",")} \\
+        --runThreadN $task.cpus \\
+        --outFileNamePrefix $prefix. \\
+        $out_sam_type \\
+        $ignore_gtf \\
+        $attrRG \\
+        $args
+
+    $mv_unsorted_bam
+
+    if [ -f ${prefix}.Unmapped.out.mate1 ]; then
+        mv ${prefix}.Unmapped.out.mate1 ${prefix}.unmapped_1.fastq
+        gzip ${prefix}.unmapped_1.fastq
+    fi
+    if [ -f ${prefix}.Unmapped.out.mate2 ]; then
+        mv ${prefix}.Unmapped.out.mate2 ${prefix}.unmapped_2.fastq
+        gzip ${prefix}.unmapped_2.fastq
+    fi
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    echo "" | gzip > ${prefix}.unmapped_1.fastq.gz
+    echo "" | gzip > ${prefix}.unmapped_2.fastq.gz
+    touch ${prefix}Xd.out.bam
+    touch ${prefix}.Log.final.out
+    touch ${prefix}.Log.out
+    touch ${prefix}.Log.progress.out
+    touch ${prefix}.sortedByCoord.out.bam
+    touch ${prefix}.toTranscriptome.out.bam
+    touch ${prefix}.Aligned.unsort.out.bam
+    touch ${prefix}.Aligned.sortedByCoord.out.bam
+    touch ${prefix}.tab
+    touch ${prefix}.SJ.out.tab
+    touch ${prefix}.ReadsPerGene.out.tab
+    touch ${prefix}.Chimeric.out.junction
+    touch ${prefix}.out.sam
+    touch ${prefix}.Signal.UniqueMultiple.str1.out.wig
+    touch ${prefix}.Signal.UniqueMultiple.str1.out.bg
+    """
+}
diff --git a/modules/local/rustar_align/genomegenerate/main.nf b/modules/local/rustar_align/genomegenerate/main.nf
@@ -0,0 +1,82 @@
+process RUSTAR_GENOMEGENERATE {
+    tag "$fasta"
+    label 'process_high'
+
+    container "ghcr.io/scverse/rustar-aligner:dev"
+
+    input:
+    tuple val(meta), path(fasta)
+    tuple val(meta2), path(gtf)
+
+    output:
+    tuple val(meta), path("star")  , emit: index
+    tuple val("${task.process}"), val('rustar-aligner'), eval("rustar-aligner --version | sed -n '1{s/^rustar-aligner //;p}'"), emit: versions_rustar, topic: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args        = task.ext.args ?: ''
+    def args_list   = args.tokenize()
+    def memory      = task.memory ? "--limitGenomeGenerateRAM ${task.memory.toBytes() - 100000000}" : ''
+    def include_gtf = gtf ? "--sjdbGTFfile $gtf" : ''
+    // Heuristic mirrors STAR_GENOMEGENERATE's gawk/samtools-faidx pipeline, but
+    // computed in Groovy so we don't need samtools+gawk in the rustar container.
+    // Approximating genome length with the on-disk fasta size is within 1-2% of
+    // the true base count and is well inside the floor() rounding of log2(len)/2-1.
+    def auto_sa_index = ''
+    if (!args_list.contains('--genomeSAindexNbases')) {
+        def genome_size = fasta.size()
+        def computed = Math.floor(Math.log(genome_size as double) / Math.log(2) / 2 - 1) as int
+        def num_bases = Math.min(14, Math.max(1, computed))
+        auto_sa_index = "--genomeSAindexNbases ${num_bases}"
+    }
+    """
+    mkdir star
+    rustar-aligner \\
+        --runMode genomeGenerate \\
+        --genomeDir star/ \\
+        --genomeFastaFiles $fasta \\
+        $include_gtf \\
+        --runThreadN $task.cpus \\
+        $auto_sa_index \\
+        $memory \\
+        $args
+    """
+
+    stub:
+    if (gtf) {
+        """
+        mkdir star
+        touch star/Genome
+        touch star/Log.out
+        touch star/SA
+        touch star/SAindex
+        touch star/chrLength.txt
+        touch star/chrName.txt
+        touch star/chrNameLength.txt
+        touch star/chrStart.txt
+        touch star/exonGeTrInfo.tab
+        touch star/exonInfo.tab
+        touch star/geneInfo.tab
+        touch star/genomeParameters.txt
+        touch star/sjdbInfo.txt
+        touch star/sjdbList.fromGTF.out.tab
+        touch star/sjdbList.out.tab
+        touch star/transcriptInfo.tab
+        """
+    } else {
+        """
+        mkdir star
+        touch star/Genome
+        touch star/Log.out
+        touch star/SA
+        touch star/SAindex
+        touch star/chrLength.txt
+        touch star/chrName.txt
+        touch star/chrNameLength.txt
+        touch star/chrStart.txt
+        touch star/genomeParameters.txt
+        """
+    }
+}
diff --git a/nextflow.config b/nextflow.config
@@ -71,6 +71,7 @@ params {
     aligner                    = 'star_salmon'
     use_sentieon_star          = false
     use_parabricks_star        = false
+    use_rustar_star            = false
     use_rustqc                 = false
     gpu_container_options      = null
     pseudo_aligner             = null
diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -447,6 +447,11 @@
                     "description": "Optionally accelerate STAR and MarkDuplicates with Parabricks",
                     "fa_icon": "fas fa-running"
                 },
+                "use_rustar_star": {
+                    "type": "boolean",
+                    "description": "Use rustar-aligner (https://github.com/scverse/rustar-aligner), a Rust port of STAR, in place of the standard STAR processes. Experimental.",
+                    "fa_icon": "fas fa-running"
+                },
                 "gpu_container_options": {
                     "type": "string",
                     "description": "Override container GPU flags for GPU tasks. Auto-detects if not set (--gpus all for Docker, --nv for Singularity/Apptainer).",
diff --git a/subworkflows/local/align_star/main.nf b/subworkflows/local/align_star/main.nf
@@ -3,6 +3,7 @@
 //
 include { SENTIEON_STARALIGN as SENTIEON_STAR_ALIGN } from '../../../modules/nf-core/sentieon/staralign/main'
 include { PARABRICKS_RNAFQ2BAM as PARABRICKS_RNA_FQ2BAM } from '../../../modules/nf-core/parabricks/rnafq2bam/main'
+include { RUSTAR_ALIGN                              } from '../../../modules/local/rustar_align/align/main'
 include { STAR_ALIGN                                } from '../../../modules/nf-core/star/align'
 include { BAM_SORT_STATS_SAMTOOLS                   } from '../../nf-core/bam_sort_stats_samtools'
 
@@ -32,6 +33,7 @@ workflow ALIGN_STAR {
     fasta_fai            // channel: [ val(meta), path(fasta), path(fai) ]
     use_sentieon_star    // boolean: whether star alignment is accelerated with Sentieon
     use_parabricks_star  // boolean: whether star alignment (and mark duplicates) is accelerated with Parabricks
+    use_rustar_star      // boolean: whether to use rustar-aligner in place of STAR
     skip_markduplicates  // boolean: whether to skip marking duplicates
 
     main:
@@ -51,6 +53,11 @@ workflow ALIGN_STAR {
         PARABRICKS_RNA_FQ2BAM(reads, fasta_fai.map { meta, fasta, _fai -> [ meta, fasta ] }, index, true, !skip_markduplicates)
         ch_star_out = PARABRICKS_RNA_FQ2BAM
 
+    } else if (use_rustar_star) {
+
+        RUSTAR_ALIGN(reads, index, gtf, star_ignore_sjdbgtf)
+        ch_star_out = RUSTAR_ALIGN
+
     } else {
 
         STAR_ALIGN(reads, index, gtf, star_ignore_sjdbgtf)
diff --git a/subworkflows/local/align_star/tests/main.extra_args.nf.test b/subworkflows/local/align_star/tests/main.extra_args.nf.test
@@ -31,6 +31,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -53,7 +54,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
diff --git a/subworkflows/local/align_star/tests/main.nf.test b/subworkflows/local/align_star/tests/main.nf.test
@@ -31,6 +31,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -53,7 +54,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
@@ -124,6 +126,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -146,7 +149,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
@@ -205,6 +209,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -227,7 +232,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
diff --git a/subworkflows/local/align_star/tests/main.parabricks.nf.test b/subworkflows/local/align_star/tests/main.parabricks.nf.test
@@ -31,6 +31,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star   = false
                 use_parabricks_star = true
+                use_rustar_star     = false
                 skip_markduplicates = false
 
                 input[0] = channel.of([
@@ -53,7 +54,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
@@ -93,6 +95,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star   = false
                 use_parabricks_star = true
+                use_rustar_star     = false
                 skip_markduplicates = false
 
                 input[0] = channel.of([
@@ -115,7 +118,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
diff --git a/subworkflows/local/align_star/tests/main.rg.nf.test b/subworkflows/local/align_star/tests/main.rg.nf.test
@@ -31,6 +31,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -53,7 +54,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
@@ -102,6 +104,7 @@ nextflow_workflow {
                 star_ignore_sjdbgtf = false
                 use_sentieon_star    = false
                 use_parabricks_star  = false
+                use_rustar_star      = false
                 skip_markduplicates  = false
 
                 input[0] = channel.of([
@@ -124,7 +127,8 @@ nextflow_workflow {
                 ])
                 input[5] = use_sentieon_star
                 input[6] = use_parabricks_star
-                input[7] = skip_markduplicates
+                input[7] = use_rustar_star
+                input[8] = skip_markduplicates
                 """
             }
         }
diff --git a/subworkflows/local/align_star/tests/main.sentieon.nf.test b/subworkflows/local/align_star/tests/main.sentieon.nf.test
diff --git a/subworkflows/local/prepare_genome_indices/main.nf b/subworkflows/local/prepare_genome_indices/main.nf
diff --git a/subworkflows/local/prepare_genome_indices/tests/main.nf.test b/subworkflows/local/prepare_genome_indices/tests/main.nf.test
diff --git a/subworkflows/local/prepare_genome_indices/tests/main.parabricks.nf.test b/subworkflows/local/prepare_genome_indices/tests/main.parabricks.nf.test
diff --git a/workflows/rnaseq/main.nf b/workflows/rnaseq/main.nf

Original file line number	Diff line number	Diff line change
`@@ -111,6 +111,7 @@ workflow NFCORE_RNASEQ {`
`111`	`111`	`params.skip_pseudo_alignment,`
`112`	`112`	`params.use_sentieon_star,`
`113`	`113`	`params.use_parabricks_star,`
	`114`	`+ params.use_rustar_star,`
`114`	`115`	`isStarIndexLegacy() ?: false`
`115`	`116`	`)`
`116`	`117`