Increase chunk size in split_reads; support GRCh37_mm10 in vcf2maf

Author: ckandoth

bin/cmo_split_reads

@@ -13,7 +13,6 @@ def chunk(fastq, platform_unit, lines_per_chunk, num_pieces):
         output_prefix = "-".join(exploded)
     while lines_per_chunk % 4 != 0:
         lines_per_chunk +=1
-    fh = io.BufferedReader(gzip.open(fastq, "rb"))
     output_file_count = 0
     output_file_lines = 0
     #these aren't relaly gz but trimgalore doesnt like files not named gz...great work trimgalore
@@ -24,17 +23,19 @@ def chunk(fastq, platform_unit, lines_per_chunk, num_pieces):
     for line in os.popen("zcat < " + fastq):
         lines.append(line)
         output_file_lines+=1
-        if output_file_lines == lines_per_chunk:
+        if output_file_lines == lines_per_chunk and int(num_pieces) > 1:
             if(output_file_count < int(num_pieces)-1):
                 output_file_lines=0
-                ofh.write("".join(lines))
+                for l in lines:
+                    ofh.write(l)
                 ofh.close()
                 lines = list()
                 output_file_count +=1
                 filename = output_prefix + "chunk{:0>3d}".format(output_file_count) +  ".fastq.gz"
                 logger.info("Opening %s and writing reads..." % (filename))
                 ofh = gzip.open(filename, "wb", 1)
-    ofh.write("".join(lines))
+    for l in lines:
+        ofh.write(l)
     ofh.close()
     return True
 
@@ -52,7 +53,7 @@ if __name__ =='__main__':
         fastqs.append(args.fastq2)
     filesize = os.path.getsize(fastqs[0])
     logger.info("Fastq1 Filesize: %sGB" % ("{:.2f}".format(float(filesize)/1000000000)))
-    num_pieces = math.ceil(float(filesize)/350000000)
+    num_pieces = math.ceil(float(filesize)/2800000000)
     logger.info("Splitting into %s pieces" % "{:.0f}".format(num_pieces))
     num_lines = sum(1 for line in os.popen("zcat " + fastqs[0]))
     lines_per_chunk = math.ceil(float(num_lines) / int(num_pieces))
@@ -62,16 +63,3 @@ if __name__ =='__main__':
         result = pool.apply_async(chunk, args=(fastq, args.platform_unit, lines_per_chunk, num_pieces, ))
     pool.close()
     pool.join()
-
-
-
-
-
-
-
-
-
-
-
-
-

bin/cmo_vcf2maf

@@ -28,9 +28,10 @@ if __name__ =='__main__':
     preparser.add_argument("--version", help="Version of tool to run", choices=cmo.util.programs['vcf2maf'].keys(), default="default")
     preparser.add_argument("--vep-release", help="Version of VEP and its cache to use", choices=cmo.util.programs['vep'].keys(), default="default")
     preparser.add_argument("--species", help="Species of variants in input", choices=["homo_sapiens","mus_musculus"], default="homo_sapiens")
-    preparser.add_argument("--ncbi-build", help="Genome build of variants in input", choices=["GRCh37","GRCh38","GRCm38"], default="GRCh37")
+    preparser.add_argument("--ncbi-build", help="Genome build of variants in input", choices=["GRCh37","GRCh38","GRCm38","GRCh37_mm10"], default="GRCh37")
     options, _ = preparser.parse_known_args()
-
+    if options.ncbi_build == 'GRCh37_mm10':
+        options.ncbi_build = 'GRCh37'
     # Figure out the path to the actual Perl script that this Python wrapper will run
     script_path = cmo.util.programs['vcf2maf'][options.version] + "vcf2maf.pl"
     # Extract arguments and their defaults, by parsing the --help output
@@ -62,6 +63,8 @@ if __name__ =='__main__':
 
     # Now run the argparse instance, which will parse and execute, or print help text if requested
     args = parser.parse_args()
+    if args.ncbi_build == 'GRCh37_mm10':
+        args.ncbi_build = 'GRCh37'
     args_dict = vars(args)
 
     # Unless user-defined, assume that sample IDs are the same in VCF and MAF