Making code Dockstore compatible

Dockstore.json

@@ -0,0 +1,23 @@
+{
+  "reference": {
+    "path": "http://hgwdev.cse.ucsc.edu/~jeltje/public_data/genome.fa.gz",
+    "class": "File"
+  },
+  "mode": "wgs",
+  "normal": {
+    "path":"https://dcc.icgc.org/api/v1/download?fn=/PCAWG/reference_data/data_for_testing/HCC1143_ds/HCC1143_BL.bam",
+    "class": "File"
+  },
+  "mutations": {
+    "path": "/tmp/muse.vcf",
+    "class": "File"
+  },
+  "known": {
+    "path": "ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606_b149_GRCh37p13/VCF/common_all_20161121.vcf.gz",
+    "class": "File"
+  },
+  "tumor": {
+    "path":"https://dcc.icgc.org/api/v1/download?fn=/PCAWG/reference_data/data_for_testing/HCC1143_ds/HCC1143.bam",
+    "class": "File"
+  }
+}

muse.cwl

@@ -0,0 +1,67 @@
+#!/usr/bin/env cwl-runner
+#
+# Author: Jeltje van Baren jeltje.van.baren@gmail.com
+
+cwlVersion: v1.0
+class: CommandLineTool
+baseCommand: [/opt/bin/muse.py, -O, muse.vcf, -w, ./, --muse, MuSEv1.0rc]
+
+doc: "Runs MuSEv1.0rc SNP caller on split chromosomes (MuSE call) then creates final calls in VCF format (MuSE sump)"
+
+hints:
+  DockerRequirement:
+    dockerPull: quay.io/opengenomics/muse
+
+#requirements:
+#  - class: InlineJavascriptRequirement
+
+inputs:
+  tumor:
+    type: File
+    doc: |
+      tumor bam file
+    inputBinding:
+      prefix: --tumor-bam
+    secondaryFiles:
+      - .bai
+  normal:
+    type: File
+    doc: |
+      normal bam file
+    inputBinding:
+      prefix: --normal-bam
+    secondaryFiles:
+      - .bai
+  reference:
+    type: File
+    doc: |
+     Reference sequence file, can be gzipped.
+    inputBinding:
+      prefix: -f
+  known:
+    type: File
+    doc: |
+      dbSNP vcf file (will be bgzip compressed and tabix indexed). Can be gzipped.
+    inputBinding:
+      prefix: -D
+  mode:
+    type: {"type": "enum", "name": "Mode", "symbols": ["wgs", "wxs"]}
+    doc: |
+      Input is whole genome or exome {wgs,wxs}
+    inputBinding:
+      prefix: --mode
+  ncpus:
+    type: int?
+    doc: |
+      number of cpus (8)
+    inputBinding:
+      position: 2
+      prefix: --cpus
+
+outputs:
+  mutations:
+    type: File
+    outputBinding:
+      glob: muse.vcf
+
+

muse.py

@@ -18,6 +18,15 @@ def which(cmd):
     if len(res) == 0: return None
     return res
 
+def gunzip(infile, outfile):
+    cmd = ' '.join(['zcat', infile])
+    with open(outfile, 'w') as outF:
+        p = subprocess.Popen(cmd, shell=True, stdout=outF, stderr=subprocess.PIPE)
+    stdout,stderr =  p.communicate()
+    if len(stderr):
+        print "unzip command failed:", stderr
+        raise Exception("unzip failed")
+
 def fai_chunk(path, blocksize):
     seq_map = {}
     with open( path ) as handle:
@@ -77,8 +86,13 @@ def run_muse(args):
         args.muse = which(args.muse)
 
     workdir = os.path.abspath(tempfile.mkdtemp(dir=args.workdir, prefix="muse_work_"))
+    if args.f.endswith('.gz'):
+        new_ref = os.path.join(workdir, "ref_genome.fasta")
+        gunzip(args.f, new_ref)
+        subprocess.check_call( ["/usr/bin/samtools", "faidx", new_ref] )
+        args.f = new_ref
 
-    if not os.path.exists(args.f + ".fai"):
+    elif not os.path.exists(args.f + ".fai"):
         new_ref = os.path.join(workdir, "ref_genome.fasta")
         os.symlink(os.path.abspath(args.f),new_ref)
         subprocess.check_call( ["/usr/bin/samtools", "faidx", new_ref] )
@@ -132,11 +146,14 @@ def run_muse(args):
             first = False
             if not args.no_clean:
                 os.unlink(out)
-
     dbsnp_file = None
     if args.D:
         new_dbsnp = os.path.join(workdir, "db_snp.vcf")
-        os.symlink(args.D,new_dbsnp)
+        if args.D.endswith('.gz'):
+            print "unzipping SNP file..."
+            gunzip(args.D, new_dbsnp)
+        else:
+            os.symlink(args.D,new_dbsnp)
         subprocess.check_call( ["/usr/bin/bgzip", new_dbsnp] )
         subprocess.check_call( ["/usr/bin/tabix", "-p", "vcf", new_dbsnp + ".gz" ])
         dbsnp_file = new_dbsnp + ".gz"