def getRawCoverageLargeMode(self, chrom, start, end, largeMode=10):
"""Retrieve an array with the genome coverage."""
out = np.zeros(end-start)
infile = BigWigFile(self.path)
wigs = infile.fetch(chrom, start, end)
for wig in wigs:
out[wig[0]-start:wig[1]-start] = wig[2]
infile.close()
return out[::largeMode]
def coverage_from_bigwig(self, bigwig_file, stepsize=100):
"""Return list of arrays describing the coverage of each genomicRegions from <bigwig_file>.
*Keyword arguments:*
- bigwig_file -- path to bigwig file
- stepsize -- used stepsize
*Output:*
Class variable <coverage>: a list where the elements correspond to the GenomicRegion. The list elements give
the number of reads falling into the GenomicRegion.
"""
if platform == "darwin" or "http" in bigwig_file:
self.coverage = []
# mp_input = []
for gr in self.genomicRegions:
# print(gr)
steps = int(abs(gr.final-gr.initial)/stepsize)
cmd = ["bigWigSummary",bigwig_file,gr.chrom,str(gr.initial-stepsize),str(gr.final-stepsize),str(steps)]
# print(" ".join(cmd))
try:
output = subprocess.check_output(cmd, shell=False, stderr=subprocess.STDOUT)
# print(output)
ds = [0 if "n/a" in x else float(x) for x in output.strip().split()]
self.coverage.append( np.array(ds) )
except:
continue
### Linux platform
else:
# print("\tUsing ngslib on linux system...")
from ngslib import BigWigFile
self.coverage = []
bwf = BigWigFile(bigwig_file)
for gr in self.genomicRegions:
depth = bwf.pileup(gr.chrom, max(0,int(gr.initial-stepsize/2)),
max(1,int(gr.final+stepsize/2)))
ds = [depth[d] for d in range(0, gr.final-gr.initial, stepsize)]
self.coverage.append( np.array(ds) )
bwf.close()
def coverage_from_bigwig(self, bigwig_file, stepsize=100):
"""Return list of arrays describing the coverage of each genomicRegions from <bigwig_file>.
*Keyword arguments:*
- bigwig_file -- path to bigwig file
- stepsize -- used stepsize
*Output:*
Class variable <coverage>: a list where the elements correspond to the GenomicRegion. The list elements give
the number of reads falling into the GenomicRegion.
"""
try:
from ngslib import BigWigFile
self.coverage = []
bwf = BigWigFile(bigwig_file)
for gr in self.genomicRegions:
depth = bwf.pileup(gr.chrom,
max(0, int(gr.initial - stepsize / 2)),
max(1, int(gr.final + stepsize / 2)))
ds = [
depth[d] for d in range(0, gr.final - gr.initial, stepsize)
]
self.coverage.append(np.array(ds))
bwf.close()
except ImportError, e:
import pyBigWig
self.coverage = []
bwf = pyBigWig.open(bigwig_file)
for gr in self.genomicRegions:
steps = int(len(gr) / stepsize)
ds = bwf.stats(gr.chrom,
gr.initial,
gr.final,
type="mean",
nBins=steps)
ds = [x if x else 0 for x in ds]
self.coverage.append(np.array(ds))
bwf.close()
def coverage_from_bigwig(self, bigwig_file, stepsize=100):
"""Return list of arrays describing the coverage of each genomicRegions from <bigwig_file>.
*Keyword arguments:*
- bigwig_file -- path to bigwig file
- stepsize -- used stepsize
*Output:*
Class variable <coverage>: a list where the elements correspond to the GenomicRegion. The list elements give
the number of reads falling into the GenomicRegion.
"""
try:
from ngslib import BigWigFile
self.coverage = []
bwf = BigWigFile(bigwig_file)
for gr in self.genomicRegions:
depth = bwf.pileup(gr.chrom, max(0, int(gr.initial - stepsize / 2)),
max(1, int(gr.final + stepsize / 2)))
ds = [depth[d] for d in range(0, gr.final - gr.initial, stepsize)]
self.coverage.append(np.array(ds))
bwf.close()
except ImportError, e:
import pyBigWig
self.coverage = []
bwf = pyBigWig.open(bigwig_file)
for gr in self.genomicRegions:
steps = int(len(gr) / stepsize)
ds = bwf.stats(gr.chrom, gr.initial, gr.final, type="mean", nBins=steps)
ds = [ x if x else 0 for x in ds ]
self.coverage.append( np.array(ds) )
bwf.close()
class BigWigDatasource(Datasource):
"""
A datasource derived from a BigWig file. For variants spanning a genomic range (i.e. non SNVs),
the median of values from the BigWig are returned.
"""
def __init__(self, src_file, title='', version=None):
# only necessary to import ngslib if instance of BigWigDatasource is created
# This should not run on OS X machines
from ngslib import BigWigFile
super(BigWigDatasource, self).__init__(src_file, title=title, version=version)
self.output_headers = [title + '_score']
self.bigwig_fh = BigWigFile(src_file)
self.has_chr = True if self.bigwig_fh.chroms[0].startswith('chr') else False
def annotate_mutation(self, mutation):
if self.has_chr and not mutation.chr.startswith('chr'):
chrn = 'chr' + mutation.chr
else:
chrn = mutation.chr
variant_start, variant_end = int(mutation.start) - 1, int(mutation.end) #start - 1 because bigwig format is zero-based coords
scores = [r[2] for r in self.bigwig_fh.fetch(chrom=chrn, start=variant_start, stop=variant_end)]
if not scores:
final_score = None
elif len(scores) == 1:
final_score = scores[0]
else:
final_score = np.median(scores)
mutation.createAnnotation(self.output_headers[0], final_score, annotationSource=self.title)
return mutation
def close(self):
self.bigwig_fh.close()
def phastCons46way_score(self, stepsize=100):
"""Load the phastCons46way bigwig files to fetch the scores as coverage.
*Keyword arguments:*
- stepsize -- used stepsize
"""
self.coverage = []
phastCons46way_dir = "/data/phastCons46way/"
for gr in self.genomicRegions:
bwf = BigWigFile(os.path.join(phastCons46way_dir, gr.chrom+".phastCons46way.bw"))
depth = bwf.pileup(gr.chrom, gr.initial-stepsize/2, gr.final+stepsize/2)
ds = []
for i in range(0, gr.final-gr.initial):
d = [ depth[j] for j in range(i,i+stepsize) ]
ds.append(sum(d)/len(d))
if gr.orientation == "-":
self.coverage.append( np.array(list(reversed(ds))) )
else:
self.coverage.append( np.array(ds) )
bwf.close()
def getChromSizesNGSLIB(self):
infile = BigWigFile(self.path)
out = infile.chromSizes()
out = dict(zip(out[0], out[1]))
infile.close()
return out
