Create a site profile vector showing the average signal accumulated from a
bigwig file around the center of each interval from a BED file.
Output is the average signal value at that relative position across the
intervals.
usage: %prog bigwig_file.bw padding < bed_file.bed
"""
import sys
from numpy import *
from bx.intervals.io import GenomicIntervalReader
from bx.bbi.bigwig_file import BigWigFile
bw = BigWigFile( open( sys.argv[1] ) )
padding = int( sys.argv[2] )
totals = zeros( padding*2, dtype=float64 )
valid = zeros( padding*2, dtype=int32 )
for interval in GenomicIntervalReader( sys.stdin ):
center = floor( ( interval.start + interval.end ) / 2 )
values = bw.get_as_array( interval.chrom, center - padding, center + padding )
# Determine which positions had data and mask the rest for totalling
invalid = isnan( values )
values[ invalid ] = 0
totals += values
valid += ( ~ invalid )
savetxt( sys.stdout, totals/valid )
def getRegionsAndGroups(regionsFileName, onlyMultiplesOf=1):
# reads a bed file containing the position
# of genomic intervals
# In case is hash sign '#' is found in the
# file, this is considered as a delimiter
# to split the heatmap into groups
regions = []
regionsDict = OrderedDict()
regionGroups = [(0, '')]
prevInterval = None
duplicates = 0
totalIntervals = 0
includedIntervals = 0
# drop some lines
for ginterval in GenomicIntervalReader(
open(regionsFileName, 'r').readlines()):
totalIntervals += 1
if ginterval.__str__()[0] == '#':
if includedIntervals > 1 and includedIntervals - regionGroups[-1][
0] > 1:
label = ginterval.__str__()[1:]
newLabel = label
if label in regionsDict.keys():
# loop to find a unique label name
i = 0
while True:
i += 1
newLabel = label + "_r" + str(i)
if newLabel not in regionsDict.keys():
break
regionsDict[newLabel] = regions[:]
regions = []
continue
# if the list of regions is to big, only consider a fraction of the data
if totalIntervals % onlyMultiplesOf != 0:
continue
# skip regions that have the same position as the previous.
# This assumes that the regions file given is sorted
if prevInterval and prevInterval.chrom == ginterval.chrom and \
prevInterval.start == ginterval.start and \
prevInterval.end == ginterval.end:
if args.verbose:
print "Gene in same region already included: %s %s:%s-%s. Skipping" % (
ginterval.fields[3], ginterval.chrom, ginterval.start,
ginterval.end)
duplicates += 1
continue
else:
prevInterval = ginterval
regions.append(intervalWrapper(ginterval))
includedIntervals += 1
if len(regions):
regionsDict[args.regionsLabel] = regions
if args.verbose:
print "%d (%.2f) regions covering the exact same interval were found" % \
(duplicates,
float(duplicates) *100 / totalIntervals)
return regionsDict
def main():
allchroms = False
options, args = doc_optparse.parse(__doc__)
try:
chr_col_1, start_col_1, end_col_1, strand_col_1 = parse_cols_arg(
options.cols1)
lengths = options.lengths
if options.all:
allchroms = True
in_fname, out_fname = args
except:
doc_optparse.exception()
g1 = NiceReaderWrapper(fileinput.FileInput(in_fname),
chrom_col=chr_col_1,
start_col=start_col_1,
end_col=end_col_1,
strand_col=strand_col_1,
fix_strand=True)
lens = dict()
chroms = list()
# dbfile is used to determine the length of each chromosome. The lengths
# are added to the lens dict and passed copmlement operation code in bx.
dbfile = fileinput.FileInput(lengths)
if dbfile:
if not allchroms:
try:
for line in dbfile:
fields = line.split("\t")
lens[fields[0]] = int(fields[1])
except:
# assume LEN doesn't exist or is corrupt somehow
pass
elif allchroms:
try:
for line in dbfile:
fields = line.split("\t")
end = int(fields[1])
chroms.append("\t".join([fields[0], "0", str(end)]))
except:
pass
# Safety...if the dbfile didn't exist and we're on allchroms, then
# default to generic complement
if allchroms and len(chroms) == 0:
allchroms = False
if allchroms:
chromReader = GenomicIntervalReader(chroms)
generator = subtract([chromReader, g1])
else:
generator = complement(g1, lens)
out_file = open(out_fname, "w")
try:
for interval in generator:
if type(interval) is GenomicInterval:
out_file.write("%s\n" % "\t".join(interval))
else:
out_file.write("%s\n" % interval)
except ParseError, exc:
out_file.close()
fail("Invalid file format: %s" % str(exc))
