def __init__(self, location, verbose=True):
if os.path.isdir(location):
self.genome = {}
filenames = os.listdir(location)
count = 0
for f in filenames:
count += 1
self.genome.update(
{f.split('.')[0]: self.readFasta(str(location + f))})
if verbose:
functions.showProgress(count, len(filenames),
'Loading genome')
if verbose: print('')
elif os.path.isfile(location):
self.genome = {}
fc = open(location, 'r')
line = fc.readline()
header, *_ = line.split()
tmp_name = header.lstrip('>')
seq = ''
line = fc.readline()
while line:
if line[0] == '>':
self.genome[tmp_name] = seq
seq = ''
header, *_ = line.split()
tmp_name = header.lstrip('>')
else:
line = line.replace('\n', '')
seq = seq + line.upper()
line = fc.readline()
self.genome[tmp_name] = seq
fc.close()
else:
sys.stderr.write('No file or directory. Exit')
def main(parclipfile,
gfffile,
upstream,
downstream,
sense,
minSize,
maxSize,
verbose,
xbins,
ybins,
vstring=''):
anno = gff.GFF(gfffile)
anno.filterSize(minSize, maxSize)
totalsize = upstream + maxSize + 1 + downstream
anno.sort2size()
pc = ParclipSiteContainer.from_file(parclipfile)
mat = []
annosize = []
for g in range(anno.size()):
tmp = [-1] * totalsize
if verbose:
functions.showProgress(g, (anno.size() - 1), vstring)
if anno.strand[g] == '+':
values = pc.getValues(anno.chr[g], anno.start[g], anno.strand[g],
sense, upstream,
(anno.stop[g] - anno.start[g]) + downstream)
else:
values = pc.getValues(anno.chr[g], anno.stop[g], anno.strand[g],
sense, upstream,
(anno.stop[g] - anno.start[g]) + downstream)
if values is not None:
tmp[0:(len(values) - 1)] = values
mat.append(functions.shrinkValues(tmp, xbins))
annosize.append(anno.stop[g] - anno.start[g])
smat = []
sannosize = []
if ybins >= anno.size():
print('Warning: --ybins >= entries in ' + gfffile)
ybins = anno.size()
ystep = round(anno.size() / ybins)
ystart = 0
ystop = ystep
while ystop < anno.size():
tmp = [0] * xbins
for i in range(xbins):
count = 0
tmpanno = 0
for j in range(ystart, ystop):
tmp[i] += mat[j][i] # [row][col]
tmpanno += annosize[j]
count += 1
tmp[i] = tmp[i] / count
tmpanno = tmpanno / count
smat.append(tmp)
sannosize.append(tmpanno)
ystart = ystop
ystop += ystep
return smat, sannosize
if verbose:
print()
def precalculateDistributions(minN, maxN, par, verbose):
distributions = [0] * (maxN + 1)
for i in range(minN, (maxN + 1)):
tmp_probabilities = []
for j in range((i + 1)):
tmp_probabilities.append(prob_bg(j, i, par))
distributions[i] = tmp_probabilities
if verbose:
functions.showProgress(i + 1, (maxN + 1),
'precalculating probability distributions')
if verbose: print('')
return (distributions)
def processPileup(file_pileup, verbose, reference='T', mutation='C'):
pseudocount = 1
maxr = 1000 #upper limit for sequencing depth which is used for parameter estimation
reference_reverse = functions.makeReverseComplement(reference)
mutation_reverse = functions.makeReverseComplement(mutation)
lines = 0
for line in open(file_pileup):
lines += 1
mr_list_neg = [[0]] #initializing empty matrices for parameter estimation
mr_list_signal = [[0]]
for i in range(1, maxr):
mr_list_neg.append([pseudocount] * (i + 1))
mr_list_signal.append([pseudocount] * (i + 1))
file_pileup = open(file_pileup, 'r')
line = file_pileup.readline()
count = percent_old = percent_new = 0
if verbose: functions.showProgress(count, lines, 'Processing Pileup')
while (line):
count += 1
split = line.split('\t')
if split[2] == reference:
tmp_counts = functions.getCounts(split[4], forward=True)
counts = [tmp_counts[0], tmp_counts[1][mutation]]
if counts[0] < maxr:
mr_list_signal[counts[0]][counts[1]] += 1
if split[2] == reference_reverse:
tmp_counts = functions.getCounts(split[4], forward=False)
counts = [tmp_counts[0], tmp_counts[1][mutation_reverse]]
if counts[0] < maxr:
mr_list_signal[counts[0]][counts[1]] += 1
if split[2] != reference and split[2] != reference_reverse:
tmp_counts = functions.getCounts(split[4], forward=True)
counts = [tmp_counts[0], max(tmp_counts[1].values())]
if counts[0] < maxr:
mr_list_neg[counts[0]][counts[1]] += 1
percent_new = math.trunc((count / lines) * 100)
if (percent_new > percent_old):
if verbose:
functions.showProgress(count, lines, 'Processing Pileup')
percent_old = percent_new
line = file_pileup.readline()
if verbose: print('')
file_pileup.close()
return ([mr_list_neg, mr_list_signal])
def getCountMat(file_pileup, minCoverage, verbose):
alphabet = ['A', 'C', 'G', 'T']
translate = {'A': 0, 'C': 1, 'G': 2, 'T': 3}
if verbose:
lines = 0
with open(file_pileup) as infile:
for line in infile:
lines += 1
mat = [[0] * 4, [0] * 4, [0] * 4, [0] * 4]
with open(file_pileup) as file_pileup:
count = 0
percent_old = 0
percent_new = 0
if verbose:
functions.showProgress(count, lines, 'Processing Pileup')
for line in file_pileup:
count += 1
split = line.split('\t')
nuc = split[2].upper()
if nuc != 'N':
tmp_counts = functions.getCounts(split[4], forward=True)
if tmp_counts[0] >= minCoverage:
for c in alphabet:
if c == nuc:
mat[translate[nuc]][translate[c]] += tmp_counts[0] - tmp_counts[2]
else:
mat[translate[nuc]][translate[c]] += tmp_counts[1][c]
if verbose:
percent_new = math.trunc(count / lines * 100)
if percent_new > percent_old:
functions.showProgress(count, lines, 'Processing Pileup')
percent_old = percent_new
if verbose:
print()
return mat
def main(parclipfile, outputfile, gfffile, downstream, upstream, gene, sense, minSize,
maxSize, verbose, vstring=''):
anno = gff.GFF(gfffile)
anno.filterSize(minSize, maxSize)
pc = ParclipSiteContainer()
pc.loadFromFile(parclipfile)
with open(outputfile, 'w') as fc_out:
for g in range(anno.size()):
if verbose:
functions.showProgress(g, (anno.size() - 1), vstring)
if anno.strand[g] == '+':
values_upstream = pc.getValues(anno.chr[g], anno.start[g], anno.strand[g],
sense, upstream, gene)
values_dostream = pc.getValues(anno.chr[g], anno.stop[g], anno.strand[g],
sense, gene, downstream)
else:
values_upstream = pc.getValues(anno.chr[g], anno.stop[g], anno.strand[g],
sense, upstream, gene)
values_dostream = pc.getValues(anno.chr[g], anno.start[g], anno.strand[g],
sense, gene, downstream)
if values_upstream is not None and values_dostream is not None:
print(*chain(values_upstream, values_dostream), sep='\t', file=fc_out)
if verbose:
print()
def main(parclip, outdir, prefix, genomepath, negset, gfffile, kmer, key,
useQuantiles, verbose, args):
scriptPath = os.path.dirname(os.path.realpath(__file__))
plot_script = os.path.join(scriptPath, 'plotKmerLogOdds.R')
pc = ParclipSiteContainer.from_file(parclip)
if gfffile is not None:
pc.remove_gff_sites(gfffile)
pc.sort(by=key, ascending=False)
kmers = functions.makekmers(kmer, list('ACGT'))[kmer - 1]
negfreq = loadNegTable(negset)
with EfficientGenome(genomepath) as genomeseq:
allfreqs = []
fileprefix = '%s_logodds_%smer_sort_%s' % (prefix, kmer, key)
if useQuantiles:
fileprefix = fileprefix + '_quantiles'
allfreqs.append(
getkmerLogs(pc, genomeseq, negfreq, kmers, 0, 1000, 15))
quantiles = [
0.01, 0.02, 0.03, 0.04, 0.05, 0.1, 0.125, 0.15, 0.175, 0.2,
0.225, 0.25, 0.275, 0.3, 0.325, 0.35, 0.375, 0.4, 0.45, 0.5,
0.55, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9
]
count = 1
stop = 1000
for q in quantiles:
if verbose:
functions.showProgress(
count, len(quantiles),
'Getting kmer log-odds from quantiles...')
old_stop = stop
start = functions.getQuantileIndex(len(pc), q) - 500
stop = functions.getQuantileIndex(len(pc), q) + 500
if start < 0:
start = 0
if stop > len(pc) - 2:
break
count = count + 1
if (stop - 500) < old_stop:
msg_pat = 'Bin %s and %s are overlapping by %s sites!'
# TODO 2x quantiles[count - 2] is probably a bug
msg = msg_pat % (quantiles[count - 2],
quantiles[count - 2], old_stop -
(stop - 500))
print(msg, file=sys.stderr)
allfreqs.append(
getkmerLogs(pc, genomeseq, negfreq, kmers, start, stop,
15))
else:
maxsize = 50000
stepsize = 1000
start = 0
stop = 1000
run = True
while run:
if stop > len(pc) - 2 or stop > maxsize:
print()
print('STOP at: %s' % +stop)
run = False
break
if verbose:
functions.showProgress(
stop, maxsize, 'Getting kmer log-odds from bins...')
allfreqs.append(
getkmerLogs(pc, genomeseq, negfreq, kmers, start, stop,
15))
start = stop
stop = stop + stepsize
table_file = os.path.join(outdir, fileprefix + '.table')
pdf_file = os.path.join(outdir, fileprefix + '.pdf')
sortAndSave(allfreqs, table_file, kmers)
cmd = [
'R',
'-q',
'--slave',
'-f %r' % plot_script,
'--args',
'%r' % table_file,
'%r' % pdf_file,
]
execute(cmd)
if not args.keep_tmp_files:
os.remove(table_file)
def findPvalueParclipInPileup(pileup,
output_file,
mincov,
maxcov,
probabilities,
verbose,
reference='T',
mutation='C',
maxPvalue=0.001,
SNPlikely=False):
reference_reverse = functions.makeReverseComplement(reference)
mutation_reverse = functions.makeReverseComplement(mutation)
found_sites = 0
lines = 0
with open(pileup) as handle:
for line in handle:
lines += 1
with open(pileup) as file_pileup, open(output_file, 'w') as file_table:
header = list(PC_MANDATORY_FIELDS) + ['p_value']
print(*header, sep='\t', file=file_table)
line = file_pileup.readline()
linecount = 0
percent_old = 0
percent_new = 0
if verbose:
functions.showProgress(linecount, lines, 'Processing Pileup')
while (line):
linecount += 1
split = line.split('\t')
counts = [0, 0]
pvalue = 1
if split[2] == reference:
tmp_counts = functions.getCounts(split[4], forward=True)
counts = [tmp_counts[0], tmp_counts[1][mutation]]
if counts[0] > mincov and counts[1] > 0:
if counts[0] > 500:
pvalue = getPvalue(
round((counts[1] / counts[0]) * 500), 500,
probabilities, SNPlikely)
else:
pvalue = getPvalue(counts[1], counts[0], probabilities,
SNPlikely)
if pvalue <= maxPvalue:
print(split[0],
split[1],
counts[1],
counts[0],
1 - pvalue,
'+',
0,
pvalue,
sep='\t',
file=file_table)
found_sites += 1
if split[2] == reference_reverse:
tmp_counts = functions.getCounts(split[4], forward=False)
counts = [tmp_counts[0], tmp_counts[1][mutation_reverse]]
if counts[0] > mincov and counts[1] > 0:
if counts[0] > 500:
pvalue = getPvalue(
round((counts[1] / counts[0]) * 500), 500,
probabilities, SNPlikely)
else:
pvalue = getPvalue(counts[1], counts[0], probabilities,
SNPlikely)
if pvalue <= maxPvalue:
print(split[0],
split[1],
counts[1],
counts[0],
1 - pvalue,
'-',
0,
pvalue,
sep='\t',
file=file_table)
found_sites += 1
percent_new = math.trunc((linecount / lines) * 100)
if (percent_new > percent_old):
if verbose:
functions.showProgress(linecount, lines,
'Processing Pileup')
percent_old = percent_new
line = file_pileup.readline()
print('Found %s PAR-CLIP sites.' % found_sites)