Add BAM input to MINIMAP2_ALIGN (nf-core#6021)

* Add BAM input * Update modules/nf-core/minimap2/align/main.nf --------- Co-authored-by: Sofia Stamouli <[email protected]>
genomic-medicine-sweden · Jul 26, 2024 · a33ef94 · a33ef94
1 parent d4ef560
commit a33ef94
Show file tree

Hide file tree

Showing 3 changed files with 490 additions and 14 deletions.
diff --git a/modules/nf-core/minimap2/align/main.nf b/modules/nf-core/minimap2/align/main.nf
@@ -28,17 +28,25 @@ process MINIMAP2_ALIGN {
     script:
     def args  = task.ext.args ?: ''
     def args2 = task.ext.args2 ?: ''
+    def args3 = task.ext.args3 ?: ''
+    def args4 = task.ext.args4 ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     def bam_index = bam_index_extension ? "${prefix}.bam##idx##${prefix}.bam.${bam_index_extension} --write-index" : "${prefix}.bam"
     def bam_output = bam_format ? "-a | samtools sort -@ ${task.cpus-1} -o ${bam_index} ${args2}" : "-o ${prefix}.paf"
     def cigar_paf = cigar_paf_format && !bam_format ? "-c" : ''
     def set_cigar_bam = cigar_bam && bam_format ? "-L" : ''
+    def bam_input = "${reads.extension}".matches('sam|bam|cram')
+    def samtools_reset_fastq = bam_input ? "samtools reset --threads ${task.cpus-1} $args3 $reads | samtools fastq --threads ${task.cpus-1} $args4 |" : ''
+    def query = bam_input ? "-" : reads
+    def target = reference ?: (bam_input ? error("BAM input requires reference") : reads)
+
     """
+    $samtools_reset_fastq \\
     minimap2 \\
         $args \\
         -t $task.cpus \\
-        ${reference ?: reads} \\
-        $reads \\
+        $target \\
+        $query \\
         $cigar_paf \\
         $set_cigar_bam \\
         $bam_output
@@ -55,6 +63,9 @@ process MINIMAP2_ALIGN {
     def prefix = task.ext.prefix ?: "${meta.id}"
     def output_file = bam_format ? "${prefix}.bam" : "${prefix}.paf"
     def bam_index = bam_index_extension ? "touch ${prefix}.bam.${bam_index_extension}" : ""
+    def bam_input = "${reads.extension}".matches('sam|bam|cram')
+    def target = reference ?: (bam_input ? error("BAM input requires reference") : reads)
+
     """
     touch $output_file
     ${bam_index}

diff --git a/modules/nf-core/minimap2/align/tests/main.nf.test b/modules/nf-core/minimap2/align/tests/main.nf.test
@@ -34,7 +34,8 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    file(process.out.bam[0][1]).name,
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
                     process.out.versions
                 ).match() }
             )
@@ -67,7 +68,8 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    file(process.out.bam[0][1]).name,
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
                     file(process.out.index[0][1]).name,
                     process.out.versions
                 ).match() }
@@ -104,7 +106,8 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    file(process.out.bam[0][1]).name,
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
                     process.out.versions
                 ).match() }
             )
@@ -137,14 +140,113 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    file(process.out.bam[0][1]).name,
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
                     process.out.versions
                 ).match() }
             )
         }
 
     }
 
+    test("sarscov2 - bam, fasta, true, [], false, false") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/genome.fasta', checkIfExists: true)
+                ]
+                input[2] = true
+                input[3] = []
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
+                    process.out.versions
+                ).match() }
+            )
+        }
+
+    }
+
+    test("sarscov2 - bam, fasta, true, 'bai', false, false") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/genome.fasta', checkIfExists: true)
+                ]
+                input[2] = true
+                input[3] = 'bai'
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(
+                    bam(process.out.bam[0][1]).getHeader(),
+                    bam(process.out.bam[0][1]).getReadsMD5(),
+                    file(process.out.index[0][1]).name,
+                    process.out.versions
+                ).match() }
+            )
+        }
+
+    }
+
+    test("sarscov2 - bam, [], true, false, false") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    []
+                ]
+                input[2] = true
+                input[3] = []
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.failed }
+            )
+        }
+
+    }
+
     test("sarscov2 - fastq, fasta, true, [], false, false - stub") {
 
         options "-stub"
@@ -241,4 +343,99 @@ nextflow_process {
 
     }
 
+    test("sarscov2 - bam, fasta, true, [], false, false - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/genome.fasta', checkIfExists: true)
+                ]
+                input[2] = true
+                input[3] = []
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+    test("sarscov2 - bam, fasta, true, 'bai', false, false - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/genome.fasta', checkIfExists: true)
+                ]
+                input[2] = true
+                input[3] = 'bai'
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+    test("sarscov2 - bam, [], true, false, false - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test', single_end:true ], // meta map
+                    file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/bam/test3.single_end.markduplicates.sorted.bam', checkIfExists: true)
+                ]
+                input[1] = [
+                    [ id:'test_ref' ], // meta map
+                    []
+                ]
+                input[2] = true
+                input[3] = []
+                input[4] = false
+                input[5] = false
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.failed }
+            )
+        }
+
+    }
+
 }