def isCoding(chr, start, end, cds_gf):
chr = chrFill(chr)
if isinstance(cds_gf, str):
gf = tabix.Tabix(cds_gf)
else:
gf = cds_gf
regs = gf.fetch('%s:%i-%i' % (chr, start, end))
regs = [reg for reg in regs]
return len(regs) >= 1
def getGeneFromCoords(
chr,
ex,
genesFile='/net/crate-04/data/burge/alexrson/finalAnalyses'
'/long_short_exons/genelines.sorted.gff.gz'):
start_q, end_q = min(ex), max(ex)
chr = chrFill(chr)
if isinstance(genesFile, str):
gf = tabix.Tabix(genesFile)
else:
gf = genesFile
regs = gf.fetch('%s:%i-%i' % (chr, start_q, end_q))
regs = [reg for reg in regs]
genes = set()
trans2gene = {}
found_trans = False
found_CDS = False
for reg in regs:
regl = reg.split('\t')
attri = readAttributes.readAttributesIntoDict(regl[8])
start_exon, end_exon = map(int, regl[3:5])
if regl[2] == 'exon':
if start_exon == start_q or start_exon == end_q or \
end_exon == start_q or end_exon == end_q:
found_trans = attri['Parent'][0]
elif regl[2] == 'CDS':
if start_exon == start_q or start_exon == end_q or \
end_exon == start_q or end_exon == end_q:
found_CDS = attri['Parent'][0]
elif regl[2] == 'mRNA':
gene = attri['Parent'][0]
trans = attri['ID'][0]
trans2gene[trans] = gene
genes.add(tuple([regl[1], gene]))
elif regl[2] == 'gene':
gene = attri['ID'][0]
genes.add(tuple([regl[1], gene]))
if found_CDS:
return trans2gene[found_CDS]
if found_trans:
return trans2gene[found_trans]
if len(genes) == 1:
return genes.pop()[1]
for esp, gene in genes:
if esp in ['protein_coding', 'rkb', 'liana']:
return gene
for esp, gene in genes:
if esp in ['ucsc.knownGene-kgXref-ensGene']:
return gene
if not genes:
print regs, chr, ex
return None
return genes.pop()[1]
chrm = 'chr%s' % (chrm)
gene = attr.split('"')[1]
ref[(chrm, start)].add(gene)
ref[(chrm, stop)].add(gene)
return ref
if __name__ == '__main__':
reffiles = sys.argv[1].split(',')
indir = sys.argv[2]
gene2go, gene2name = get_go_bits_from_david()
known2ens = get_known2ens()
ref = get_ref(reffiles)
tab = tabix.Tabix(
'/net/afterthefact/data/jmerkin/Mus_musculus.NCBIM37.67.gtf.gz')
types = []
faileds = []
conv_file = open('convert_allevents_id2gene', 'w')
for misotype in os.listdir(indir):
reffunc, this_ref = get_reffunc(misotype, known2ens, ref)
print misotype, reffunc
if reffunc is None: continue
types.append(misotype)
fi = open(
'%s/%s/Comparisons/N2ASoma_vs_N2AAxon/bayes-factors/N2ASoma_vs_N2AAxon.miso_bf'
% (
indir,
misotype,
def setUp(self):
self.tb = tabix.Tabix(EXAMPLEFILE)
import sys, tabix, optparse
p = optparse.OptionParser()
p.add_option(
'-p',
'--padded',
action='store',
dest='pad',
help=
'Program will remove this amount of padding when checking for scores, but will remain in the output bed file',
default=0)
options, args = p.parse_args()
# PWM bed
bed = open(args[0], 'rU')
# PhastCons
try:
fc = tabix.Tabix(args[2])
except:
fc = tabix.Tabix('fastcons44.bed.gz')
#Output
w = open(args[1], 'w')
for line in bed:
scores = []
l1 = line.strip().split('\t')
chrom = l1[0]
#Positive strand
if int(l1[1]) > 0:
start = int(l1[1])
end = int(l1[2])
#Negative strand
else:
def main():
reffile = sys.argv[1]
indir = sys.argv[2]
minbf = float(sys.argv[3])
minpsi = .25
#gene2go = get_go_bits()
gene2go, gene2name = get_go_bits_from_david()
go2gene = defaultdict(set)
these, alls = [{} for _ in xrange(2)]
faileds, types = [[] for _ in xrange(2)]
go_fore, go_back = [defaultdict(int) for _ in xrange(2)]
tab = tabix.Tabix(
'/net/afterthefact/data/jmerkin/Mus_musculus.NCBIM37.67.gtf.gz')
known2ens = get_known2ens()
ref = get_ref(reffile)
conv_file = open('convert_allevents_id2gene', 'w')
for misotype in os.listdir(indir):
reffunc, this_ref = get_reffunc(misotype, known2ens, ref)
if reffunc is None: continue
types.append(misotype)
fi = open(
'%s/Comparisons/N2ASoma_vs_N2AAxon/bayes-factors/N2ASoma_vs_N2AAxon.miso_bf'
% (misotype, ), 'r')
line = fi.readline()
conv_file.write(line)
#import code ; code.interact(local=locals())
for line in fi:
line = line.split('\t')
gene = reffunc(this_ref, line, tab)
if gene:
if float(line[8]) > minbf and abs(float(line[7])) > minpsi:
these[gene] = True
alls[gene] = True
conv_file.write('%s\t%s\t%s' %
(misotype, gene, '\t'.join(line)))
else:
faileds.append(misotype)
conv_file.close()
types = '.'.join(types)
for gene in alls:
for go in gene2go[gene]:
go_back[go] += 1
for gene in these:
for go in gene2go[gene]:
go_fore[go] += 1
go2gene[go].add('%s:%s' % (gene, gene2name[gene]))
fore = len(these)
back = len(alls)
scoreds, folds = [[] for _ in xrange(2)]
for go in go_back:
back_with = go_back[go]
fore_with = go_fore[go]
if min(back_with, fore_with) < 2: continue
#fore_with = max(fore_with-1 , 0)
back_without = back - back_with
fore_without = fore - fore_with
try:
fold = float(fore_with * back) / float(back_with * fore)
except:
fold = 'NA'
table = [
#[back_with, back_without],
#[fore_with, fore_without]
#[back_with, fore_with],
#[back_without, fore_without]
[fore_with, fore_without],
[back_with, back_without]
]
#print table
table = np.array(table)
pval = fisher_exact(table, alternative='greater')[1]
scoreds.append((pval, go, fold))
scoreds.sort(key=lambda xx: xx[0])
scores, names, folds = zip(*scoreds)
names = np.array(names)
scores = np.array(scores)
nscores = scores.shape[0]
folds = np.array(folds)
bonferroni = np.minimum(scores * float(nscores), 1.)
benjamini = []
oldp, knum, store = 0, 0, 0
for ii in scores:
benjamini.append(ii * nscores / (nscores - knum))
#benjamini.append(ii * (nscores - knum) / nscores )
store += 1
if oldp == ii:
# to handle ties. count number of tied scores, then add them later
pass
else:
knum += store
store = 0
oldp = ii
benjamini = np.minimum(np.array(benjamini), 1.)
print faileds
print 'failed', len(faileds)
print 'these', len(these)
print 'all', len(alls)
nscores = scores.shape[0]
iis = np.arange(nscores) + 1
Q = 0.05
Qs = iis * Q / nscores
def test_ben(pv, ii, ll, Q=0.05):
if pv < ii * Q / ll:
return True
else:
return False
f_end = 'bf%s_psi%s_%s' % (minbf, minpsi, types)
outf = open('go_analyses_%s' % (f_end), 'w')
outf.write('term\tp-value\tbenjamini\tfdr\tfold_enrich\tgenes\n')
passed = False
for go, pvalue, bonf, benj, fold, ind, qv in reversed(
zip(names, scores, bonferroni, benjamini, folds, iis, Qs)):
if fold < 1: continue
#if benj > 0.05: break
#print pvalue, qv
if passed or pvalue < qv: #test_ben(pvalue, ind, nscores, Q=Q):
#print pvalue, benj, bonf
line = '\t'.join(
map(str, [go, pvalue, benj, qv, fold, ';'.join(go2gene[go])]))
outf.write(line)
outf.write('\n')
passed = True
outf.close()
fout = open('genes_sigdif_%s' % (f_end), 'w')
for gene in these:
fout.write(gene)
fout.write('\n')
fout.close()
fout = open('genes_all_%s' % (f_end), 'w')
for gene in alls:
fout.write(gene)
fout.write('\n')
fout.close()
'''
Appends the average conservation score of a region to each entry in a bed file
usage:
python conservationBed.py [OPTIONS] bedfile outputfile <conservationfile>
'''
import sys, tabix, optparse
p = optparse.OptionParser()
p.add_option('-p', '--padded',action = 'store', dest = 'pad', help = 'Program will remove this amount of padding when checking for scores, but will remain in the output bed file', default = 0)
options, args = p.parse_args()
# PWM bed
bed = open(args[0], 'rU')
# PhastCons
try:
fc = tabix.Tabix(args[2])
except:
fc = tabix.Tabix('/gen_local/hsuj/ref/PWM/fastcons44.bed.gz')
#Output
w = open(args[1], 'w')
def main():
for line in bed:
scores = []
l1 = line.strip().split('\t')
chrom = l1[0]
start = int(l1[1])
end = int(l1[2])
start += int(options.pad)
end -= int(options.pad)