Merge pull request #116 from icgc-argo/bam-merge-sort-markdup.0.1.11.0

Bam merge sort markdup.0.1.11.0
icgc-argo-workflows · Nov 10, 2020 · a3d1200 · a3d1200
2 parents 3bcce13 + f792826
commit a3d1200
Show file tree

Hide file tree

Showing 4 changed files with 31 additions and 13 deletions.
diff --git a/tools/bam-merge-sort-markdup/bam-merge-sort-markdup.nf b/tools/bam-merge-sort-markdup/bam-merge-sort-markdup.nf
@@ -22,7 +22,7 @@
  */
 
 nextflow.enable.dsl=2
-version = '0.1.10.0'
+version = '0.1.11.0'
 
 params.aligned_lane_bams = ""
 params.ref_genome_gz = ""

diff --git a/tools/bam-merge-sort-markdup/bam-merge-sort-markdup.py b/tools/bam-merge-sort-markdup/bam-merge-sort-markdup.py
@@ -46,7 +46,7 @@ def main():
     parser.add_argument("-n", "--cpus", dest='cpus', type=int, default=cpu_count())
     parser.add_argument("-d", "--mdup", dest='mdup', action='store_true')
     parser.add_argument("-l", "--lossy", dest='lossy', action='store_true')
-    parser.add_argument("-o", "--output-format", dest='output_format', choices=['bam', 'cram'], default=['cram'], nargs='+')
+    parser.add_argument("-o", "--output-format", dest='output_format', default='cram', choices=['bam', 'cram'])
 
     args = parser.parse_args()
 
@@ -66,21 +66,14 @@ def main():
     else:
         cram = 'samtools view -C -T %s -@ %s --write-index /dev/stdin -o %s ' % (args.reference, args.cpus, args.output_base + ".cram")
 
-    tee = 'tee %s ' % (args.output_base + ".bam")
-    bai = 'samtools index -@ %s /dev/stdin %s' % (args.cpus, args.output_base + ".bam.bai")
-    bai1 = 'samtools index -@ %s %s %s ' % (args.cpus, args.output_base + ".bam", args.output_base + ".bam.bai")
+    bam = 'samtools view -b -h -@ %s --write-index /dev/stdin -o %s##idx##%s ' % (args.cpus, args.output_base + ".bam", args.output_base + ".bam.bai")
     crai1 = 'samtools index -@ %s %s %s ' % (args.cpus, args.output_base + ".cram", args.output_base + ".cram.crai")
 
     # build command
-    if "bam" in args.output_format and "cram" in args.output_format:
-        cmd.append('|'.join([merge, tee, cram]))
-        cmd.append(bai1)
-        if args.lossy: cmd.append(crai1)
-
-    elif "bam" in args.output_format and not "cram" in args.output_format:
-        cmd.append('|'.join([merge, tee, bai]))
+    if args.output_format == 'bam':
+        cmd.append('|'.join([merge, bam]))
 
-    elif not "bam" in args.output_format and "cram" in args.output_format:
+    elif args.output_format == 'cram':
         cmd.append('|'.join([merge, cram]))
         if args.lossy: cmd.append(crai1)
     else:

diff --git a/tools/bam-merge-sort-markdup/tests/test-job2.nf.json b/tools/bam-merge-sort-markdup/tests/test-job2.nf.json
@@ -0,0 +1,13 @@
+{
+  "aligned_lane_bams": [
+    "input/grch38-aligned.D0RH0_2.lane.bam",
+    "input/grch38-aligned.D0RE2_1.lane.bam",
+    "input/grch38-aligned.C0HVY_2.lane.bam"
+  ],
+  "ref_genome_gz": "reference/tiny-grch38-chr11-530001-537000.fa.gz",
+  "tempdir": "/tmp",
+  "markdup": false,
+  "output_format": "bam",
+  "cpus": 1,
+  "mem": 2
+}
diff --git a/tools/bam-merge-sort-markdup/tests/test-job3.nf.json b/tools/bam-merge-sort-markdup/tests/test-job3.nf.json
@@ -0,0 +1,12 @@
+{
+  "aligned_lane_bams": [
+    "input/grch38-aligned.D0RH0_2.lane.bam",
+    "input/grch38-aligned.D0RE2_1.lane.bam",
+    "input/grch38-aligned.C0HVY_2.lane.bam"
+  ],
+  "ref_genome_gz": "reference/tiny-grch38-chr11-530001-537000.fa.gz",
+  "tempdir": "/tmp",
+  "output_format": "bam",
+  "cpus": 1,
+  "mem": 2
+}