def test_bw():
b = BigWig("libBigWig/test/test.bw")
assert repr(b) == "BigWig('libBigWig/test/test.bw')"
intervals = list(b("1", 0, 99))
assert intervals[0] == Interval(chrom='1', start=0, end=1, value=0.10000000149011612)
assert intervals[1] == Interval(chrom='1', start=1, end=2, value=0.20000000298023224)
assert intervals[2] == Interval(chrom='1', start=2, end=3, value=0.30000001192092896)
# default is to include all values
vals = b.values("1", 0, 9)
exp = array('f', [0.10000000149011612, 0.20000000298023224, 0.30000001192092896, nan, nan, nan, nan, nan, nan])
arr_equal(vals, exp)
vals = b.values("1", 0, 9, False)
exp = array('f', [0.10000000149011612, 0.20000000298023224, 0.30000001192092896])
arr_equal(vals, exp)
v = b.stats("1", 0, 9)
assert v == 0.2000000054637591
v = b.stats("1", 0, 9, stat="stdev")
assert v == 0.10000000521540645
v = b.stats("1", 0, 4, stat="coverage")
assert v == 0.75
v = b.stats("1", 0, 4, stat="coverage", nBins=2)
assert v == array('d', [1.0, 0.5])
b.close()
def read_gerp(region, sends, path='/scratch/ucgd/lustre/u1021864/serial/hg19.gerp.bw'):
gerp = BigWig(path)
exongerp=[]
chrom, se = region.split(":")
s, e = map(int, se.split("-"))
if not chrom.startswith("chr"):
chrom = "chr" + chrom
for key in sends:
for (exs, exe) in zip(sends[key][0], sends[key][1]):
exongerp.extend(np.frombuffer(gerp.values(chrom, int(exs)-1, int(exe)), dtype='f'))
return np.frombuffer(gerp.values(chrom, int(s)-1, int(e)), dtype='f'), exongerp
def build_dnase_fc_scores(self):
path=DNASE_FOLD_COV_DIR
scores = np.zeros((len(self), len(self.samples)), dtype=float)
for sample_i, sample_name in enumerate(self.samples):
fname = "DNASE.{}.fc.signal.bigwig".format(sample_name)
b = BigWig(os.path.join(path, fname))
for region_i, region in enumerate(self.iter_regions()):
if region_i%1000000 == 0:
print "Sample %i/%i, row %i/%i" % (
sample_i+1, len(self.samples), region_i, len(self))
scores[region_i, sample_i] = b.stats(
region.contig, region.start, region.stop, 'mean')
b.close()
return pd.DataFrame(
np.nan_to_num(scores), columns=self.samples, index=self.data.index)
def test_bw():
b = BigWig("libBigWig/test/test.bw")
assert repr(b) == "BigWig('libBigWig/test/test.bw')"
intervals = list(b("1", 0, 99))
assert intervals[0] == Interval(chrom='1',
start=0,
end=1,
value=0.10000000149011612)
assert intervals[1] == Interval(chrom='1',
start=1,
end=2,
value=0.20000000298023224)
assert intervals[2] == Interval(chrom='1',
start=2,
end=3,
value=0.30000001192092896)
# default is to include all values
vals = b.values("1", 0, 9)
exp = array('f', [
0.10000000149011612, 0.20000000298023224, 0.30000001192092896, nan,
nan, nan, nan, nan, nan
])
arr_equal(vals, exp)
vals = b.values("1", 0, 9, False)
exp = array(
'f', [0.10000000149011612, 0.20000000298023224, 0.30000001192092896])
arr_equal(vals, exp)
v = b.stats("1", 0, 9)
assert v == 0.2000000054637591
v = b.stats("1", 0, 9, stat="stdev")
assert v == 0.10000000521540645
v = b.stats("1", 0, 4, stat="coverage")
assert v == 0.75
v = b.stats("1", 0, 4, stat="coverage", nBins=2)
assert v == array('d', [1.0, 0.5])
b.close()
def build_dnase_fc_scores(self):
path = DNASE_FOLD_COV_DIR
scores = np.zeros((len(self), len(self.samples)), dtype=float)
for sample_i, sample_name in enumerate(self.samples):
fname = "DNASE.{}.fc.signal.bigwig".format(sample_name)
b = BigWig(os.path.join(path, fname))
for region_i, region in enumerate(self.iter_regions()):
if region_i % 1000000 == 0:
print "Sample %i/%i, row %i/%i" % (
sample_i + 1, len(self.samples), region_i, len(self))
scores[region_i,
sample_i] = b.stats(region.contig, region.start,
region.stop, 'mean')
b.close()
return pd.DataFrame(np.nan_to_num(scores),
columns=self.samples,
index=self.data.index)
def perchrom(pli_gerp_chrom):
plis = []
gerps = []
plipath, gerppath, chrom = pli_gerp_chrom
pLI = tabix.open(plipath)
gerp = BigWig(gerppath)
gerpdict = {}
lengths = []
scores = []
rangeprev = None
for region in pLI.querys(chrom):
gerpscore, overlap = read_gerp(gerp, region)
gerps.append(float(gerpscore))
plis.append(float(region[-1]))
return gerps, plis
def perchrom(ccr_gerp_chrom):
ccrpath, gerppath, chrom = ccr_gerp_chrom
ccr = tabix.open(ccrpath)
gerp = BigWig(gerppath)
gerpdict={}; gerps=[]
for region in ccr.querys(chrom):
gene=region[3]; ranges=region[6]; pctile=float(region[-1])
gerpscore, overlap = read_gerp(gerp, region) # _ = pfam, redundant variable
gerps.append((gerpscore, overlap, ranges, gene, pctile))
sorter = itemgetter(2,3)
grouper = itemgetter(2,3)
for key, grp in groupby(sorted(gerps, key = sorter), grouper):
lengths = []; scores = []
grp = list(grp)
ranges = grp[0][2]; gene = grp[0][3]; pctile = grp[0][-1]
for i, elem in enumerate(grp):
scores.append(grp[i][0])
lengths.append(grp[i][1])
gerpscore=sum([a*b for a,b in zip(scores,lengths)])/sum(lengths)
gerpdict[key]=(gerpscore,pctile,gene,sum(lengths),ranges,chrom)
return gerpdict
scores.append(grp[i][1])
lengths.append(grp[i][-1])
famscore = sum([a * b for a, b in zip(scores, lengths)]) / sum(lengths)
gerpdict[family] = famscore
return gerpdict
#pfampath='pfam.hg19.bed' # pfam doms incl. introns
pfampath = sys.argv[
1] # from '/uufs/chpc.utah.edu/common/home/u1021864/analysis/pfam/pfam.genome.bed' # sorted by pfam name, in genome space
gerppath = sys.argv[2] # '/scratch/ucgd/lustre/u1021864/serial/hg19.gerp.bw'
ccrpath = sys.argv[
3] # '/uufs/chpc.utah.edu/common/home/u1021864/analysis/exacresiduals/gnomad10x.5syn-ccrs.bed.gz'
gerp = BigWig(gerppath)
pfams = read_pfam(pfampath)
#ccr = tabix.open(ccrpath)
ccrs = score_average(pfampath, ccrpath)
gerps = score_gerp(pfams, gerp)
#for i in ccrs:
# print i, ccrs[i]
#for i in gerps:
# print i, gerps[i]
cscores, gscores, labels = [], [], []
for pfam in ccrs:
cscores.append(ccrs[pfam])
gscores.append(gerps[pfam])
labels.append(pfam)
def test_bad_chr():
b = BigWig("libBigWig/test/test.bw")
assert b.stats("chr1", 0, 10) is None
v = b.values("chr1", 0, 10)
assert len(v) == 0, v
def test_seqs():
b = BigWig("libBigWig/test/test.bw")
assert b.chroms == [('1', 195471971), ('10', 130694993)], b.chroms
def test_bad_chr():
b = BigWig("libBigWig/test/test.bw")
assert b.stats("chr1", 0, 10) is None
v = b.values("chr1", 0, 10)
assert len(v) == 0, v