ability to specify chromosomes to process for bplib building

bioinform · Nov 21, 2015 · 8cf3cdb · 8cf3cdb
1 parent 4d162a5
commit 8cf3cdb
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 2 deletions.
diff --git a/breakseq2/breakseq_index.py b/breakseq2/breakseq_index.py
@@ -22,7 +22,7 @@ def get_seq(sv, jn_type, seq, format_version):
     return ">%s:%s:%s\n%s" % (sv.id, sv.size(), jn_type, seq)
 
 
-def generate_bplib(gff, reference, output, junction_length=DEFAULT_JUNCTION_LENGTH, format_version=DEFAULT_FORMAT_VERSION):
+def generate_bplib(gff, reference, output, junction_length=DEFAULT_JUNCTION_LENGTH, format_version=DEFAULT_FORMAT_VERSION, chromosomes=[]):
     logger = logging.getLogger(generate_bplib.__name__)
 
     if not gff or not os.path.isfile(gff):
@@ -37,6 +37,9 @@ def generate_bplib(gff, reference, output, junction_length=DEFAULT_JUNCTION_LENG
         logger.error("Insertion sequence file %s not found. Insertions will be skipped" % ins_file)
 
     for sv in SV.parse(gff, Fasta.Seqs(reference, junction_length)):
+        if chromosomes and sv.name not in chromosomes:
+            continue
+
         if sv.is_insertion() and ins_file_absent:
             logger.warn("Omitting entry %s due to missing insertion sequence file" % sv.id)
             continue

diff --git a/scripts/breakseq2_gen_bplib.py b/scripts/breakseq2_gen_bplib.py
@@ -11,6 +11,7 @@
     breakseq_index.add_options(parser)
     parser.add_argument("--reference", help="Reference FASTA", required=True)
     parser.add_argument("--output", help="Output FASTA to generate. Leave unspecified for stdout")
+    parser.add_argument("--chromosomes", nargs="+", help="List of chromosomes to process", default=[])
     parser.add_argument('--version', action='version', version='%(prog)s ' + _version.__version__)
     parser.add_argument('--log', help="Log level", default="INFO")
 
@@ -26,4 +27,4 @@
     logger = logging.getLogger(__file__)
     logger.info("Command-line: " + " ".join(sys.argv))
 
-    breakseq_index.generate_bplib(args.bplib_gff, args.reference, args.output, args.junction_length, args.format_version)
+    breakseq_index.generate_bplib(args.bplib_gff, args.reference, args.output, args.junction_length, args.format_version, args.chromosomes)