| 1 |
#!/usr/bin/env python |
| 2 |
# encoding: utf-8 |
| 3 |
""" |
| 4 |
bed_to_juncs.py |
| 5 |
|
| 6 |
Created by Cole Trapnell on 2008-09-19. |
| 7 |
Copyright (c) 2008 Cole Trapnell. All rights reserved. |
| 8 |
""" |
| 9 |
|
| 10 |
import sys |
| 11 |
import getopt |
| 12 |
|
| 13 |
|
| 14 |
help_message = ''' |
| 15 |
This script converts junctions in BED format produced by TopHat to the |
| 16 |
internal .juncs format for re-use with future runs. |
| 17 |
|
| 18 |
Usage: |
| 19 |
|
| 20 |
bed_to_juncs.py < junctions.bed |
| 21 |
''' |
| 22 |
|
| 23 |
|
| 24 |
class Usage(Exception): |
| 25 |
def __init__(self, msg): |
| 26 |
self.msg = msg |
| 27 |
|
| 28 |
|
| 29 |
def main(argv=None): |
| 30 |
if argv is None: |
| 31 |
argv = sys.argv |
| 32 |
try: |
| 33 |
try: |
| 34 |
opts, args = getopt.getopt(argv[1:], "h", ["help"]) |
| 35 |
except getopt.error, msg: |
| 36 |
raise Usage(msg) |
| 37 |
|
| 38 |
for option, value in opts: |
| 39 |
if option in ("-h", "--help"): |
| 40 |
raise Usage(help_message) |
| 41 |
|
| 42 |
line_num = 0 |
| 43 |
for line in sys.stdin.readlines(): |
| 44 |
line = line.strip() |
| 45 |
cols = line.split() |
| 46 |
line_num += 1 |
| 47 |
if len(cols) < 12: |
| 48 |
print >> sys.stderr, "Warning: malformed line %d, missing columns" % line_num |
| 49 |
print >> sys.stderr, "\t", line |
| 50 |
continue |
| 51 |
chromosome = cols[0] |
| 52 |
orientation = cols[5] |
| 53 |
block_starts = [int(x) for x in cols[11].split(",")] |
| 54 |
block_sizes = [int(x) for x in cols[10].split(",")] |
| 55 |
|
| 56 |
left_pos = int(cols[1]) + block_starts[0] + block_sizes[0] - 1 |
| 57 |
right_pos = int(cols[1]) + block_starts[1] |
| 58 |
print "%s\t%d\t%d\t%s" % (chromosome, left_pos, right_pos, orientation) |
| 59 |
|
| 60 |
|
| 61 |
except Usage, err: |
| 62 |
print >> sys.stderr, sys.argv[0].split("/")[-1] + ": " + str(err.msg) |
| 63 |
print >> sys.stderr, "\t for help use --help" |
| 64 |
return 2 |
| 65 |
|
| 66 |
|
| 67 |
if __name__ == "__main__": |
| 68 |
sys.exit(main()) |
