def main(parclipA, parclipB, outfile, width, verbose):
quantiles = [0,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0]
total = (len(quantiles)-1)*(len(quantiles)-1)
total_count = 0
if verbose:
functions.showProgress(total_count, total, 'Calculating Jaccard-Index')
fc = open(outfile, 'w')
for q1 in range(len(quantiles)-1):
a = ParclipSiteContainer()
a.loadFromFile(parclipA)
aq = getEntries(a,quantiles[q1], quantiles[q1+1])
#removeEntries(a,quantiles[q1], quantiles[q1+1])
for q2 in range(len(quantiles)-1):
b = ParclipSiteContainer()
b.loadFromFile(parclipB)
#removeEntries(b,quantiles[q2], quantiles[q2+1])
bq = getEntries(b,quantiles[q2], quantiles[q2+1])
intersect = 0
for j in range(bq.size()):
if aq.exactSearch(bq.chrs[j], bq.pos[j], bq.strand[j], width=width)[1]:
intersect += 1
jaccard = intersect/(aq.size()+bq.size()-intersect)
#print('q1: '+str(quantiles[q1])+' q2: '+str(quantiles[q2])+' '+str(round(jaccard,4)))
fc.write(str(round(jaccard,4))+'\t')
total_count += 1
if verbose:
functions.showProgress(total_count, total, 'Calculating Jaccard-Index')
fc.write('\n')
print('')
fc.close()
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 main(inputfile, outputfile):
if os.path.isfile(inputfile) == False:
print('Inputfile: '+inputfile+' does not exist')
sys.exit(-1)
sites = ParclipSiteContainer()
sites.loadFromFile(inputfile)
for i in range(sites.size()):
sites.occ[i] = sites.m[i]/sites.r[i]
sites.save2File(outputfile)
def main(input_file, output_file, q):
if not 0 <= q < 1:
print('q must lie between 0 and 1 - got %s' % q)
sys.exit(1)
sites = ParclipSiteContainer.from_file(input_file)
# dirty hack to avoid errors on empty files
occ_vals = []
for rec in sites:
occ_vals.append(rec.occupancy)
if len(occ_vals) > 0:
max_occ = functions.getQuantile(occ_vals, q)
records = []
for rec in sites:
if rec.occupancy > max_occ:
rec = rec._replace(occupancy=max_occ)
records.append(rec)
new_sites = ParclipSiteContainer(records)
new_sites.save2File(output_file)
def getEntries(sites, q1, q2):
i = 0
lower = functions.getQuantile(sites.occ,q1)
upper = functions.getQuantile(sites.occ,q2)
pc = ParclipSiteContainer()
count = 0
size = sites.size()
for i in range(sites.size()):
if sites.occ[i] > lower and sites.occ[i] <= upper:
count += 1
pc.addSite(sites.chrs[i], sites.pos[i], sites.m[i], sites.r[i], sites.result[i], sites.strand[i], sites.occ[i])
pc.getChromosomePositions()
return pc
def run():
scriptPath = os.path.dirname(os.path.realpath(__file__))
plot_script = os.path.join(scriptPath, 'plotKmerPerPosition.R')
parser = create_parser()
args = parser.parse_args()
sites = ParclipSiteContainer.from_file(args.inputfile)
if args.filterGFF != '':
sites.remove_gff_sites(args.filterGFF, args.awidth)
sites.sort(by=args.key, ascending=False)
with EfficientGenome(args.genome) as genome:
sites = sites[args.start:args.stop]
seqs = sites.get_all_sequences(genome, args.width)
prefix_fmt = '%s_kmerPerPosition_kmer%s_start%s_stop%s_width%s_sort_%s'
prefix = prefix_fmt % (args.prefix, args.kmer, args.start, args.stop,
args.width, args.key)
outfile_table = os.path.join(args.outdir, prefix + '.table')
outfile_pdf = os.path.join(args.outdir, prefix + '.pdf')
seq_len = 2 * args.width + 1
getKmerOccurences(seqs,
seq_len,
outfile_table,
kmer=(args.kmer - 1),
verbose=args.verbose)
cmd = [
'R', '-q', '--slave',
'-f %s' % plot_script, '--args', outfile_table, outfile_pdf,
args.width, 0, args.width + 1
]
execute(cmd)
if args.remove:
os.remove(outfile_table)
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 run():
scriptPath = os.path.dirname(os.path.realpath(__file__))
parser = create_parser()
args = parser.parse_args()
prefix_pat = '%s_xxmotif_start%s_stop%s_width%s_sort_%s'
file_prefix = prefix_pat % (args.prefix, args.start, args.stop, args.width,
args.key)
sites = ParclipSiteContainer.from_file(args.inputfile)
if args.filterGFF != '':
sites.remove_gff_sites(args.filterGFF, args.awidth)
sites.sort(by=args.key, ascending=False)
sites = sites[args.start:args.stop]
gen_file = os.path.join(args.outdir, file_prefix + '.fa')
with EfficientGenome(args.genome) as genome:
sites.save2Fasta(genome, gen_file, width=args.width)
cmd = [
'XXmotif',
args.outdir,
gen_file,
'--zoops',
'--merge-motif-threshold LOW',
'--max-match-positions 10',
]
if args.negSet:
cmd.append('--negSet %s' % args.negSet)
execute(cmd)
tmp_dir = os.path.join(args.outdir, 'tmp')
mini_plot_script = os.path.join(tmp_dir, 'plotDistribution.R')
mini_plot_cmd = [
'R',
'-q',
'--slave',
'-f %r' % mini_plot_script,
'--args',
'%r' % args.outdir,
]
execute(mini_plot_cmd)
plot_script = os.path.join(scriptPath, '..', 'plots', 'weblogo.R')
pwm_file = os.path.join(args.outdir, file_prefix + '.pwm')
plot_cmd = [
'R',
'-q',
'--slave',
'-f %s' % plot_script,
'--args',
pwm_file,
args.outdir,
file_prefix,
args.plotPWM,
]
if args.plotPWM > 0:
execute(plot_cmd)
if not args.keep_tmp_files:
shutil.rmtree(tmp_dir, ignore_errors=True)
def main(parclipA, parclipB, start, stop, width, anno=None, annowidth=100,
logRatio=False, verbose=False):
tmpA = ParclipSiteContainer()
dataB = ParclipSiteContainer()
tmpA.loadFromFile(parclipA)
tmpA.sort(key='occ')
dataB.loadFromFile(parclipB)
if start < 0 or stop < start or stop >= tmpA.size():
print('Bullshit start and stop indices. Come on! Concentrate!')
sys.exit()
dataA = parclipsites.ParclipSites('')
total = stop - start
count = 0
i = start
while count < total and i < (tmpA.size()-1):
if verbose:
functions.showProgress(count,total-1,'Selecting PAR-CLIP sites')
if anno == None:
dataA.addSite(tmpA.chrs[i], tmpA.pos[i], tmpA.m[i], tmpA.r[i],
tmpA.result[i], tmpA.strand[i], tmpA.occ[i])
count +=1
else:
if anno.isInside(tmpA.chrs[i], tmpA.pos[i], tmpA.strand[i],
annowidth, annowidth)[1]:
dataA.addSite(tmpA.chrs[i], tmpA.pos[i], tmpA.m[i], tmpA.r[i],
tmpA.result[i], tmpA.strand[i], tmpA.occ[i])
count +=1
i += 1
coloc = 1
count_coloc = 1
if verbose:
print('\n')
for i in range(dataA.size()):
values = dataB.getValues(dataA.chrs[i], dataA.pos[i], dataA.strand[i], True, width, width)
if values != None:
count_coloc += 1
coloc += max(values)
if verbose:
functions.showProgress(i, (dataA.size()-1), 'Collecting colocolization data')
coloc = coloc / count_coloc
if verbose:
print('')
if logRatio:
return math.log( coloc/functions.getQuantile(dataB.occ,0.5) ,2)
else:
return coloc
def _get_container(self):
pc_table = ParclipSiteContainer.from_file(TABLE_DIR)
return pc_table
def main(inputfile, outputfile, gfffile, gffmin, gffmax, takeStop, upstream,
downstream, verbose):
takeStart = True
if takeStop:
takeStart = False
sites = ParclipSiteContainer()
sites.loadFromFile(inputfile)
anno = gff.GFF(gfffile)
anno.filterSize(gffmin, gffmax)
anno.getChromosomePositions()
if anno.size() < 10:
print('Warning: Low number of annotation enries! ' + str(anno.size()))
fsites = ParclipSiteContainer()
percent_old = 0
percent_new = 0
for i in range(sites.size()):
if anno.isAround(sites.chrs[i], sites.pos[i], sites.strand[i],
takeStart, upstream, downstream)[1]:
fsites.addSite(sites.chrs[i], sites.pos[i], sites.m[i], sites.r[i],
sites.result[i], sites.strand[i], sites.occ[i])
percent_new = round(i / sites.size() * 100)
if percent_new > percent_old:
if verbose:
functions.showProgress(i, anno.size(), 'selecting sites')
percent_old = percent_new
fsites.save2File(outputfile)
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)
from mockinbird.utils import ParclipSiteContainer
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Takes PAR-CLIP sites and a genome and saves genomic sequences as fasta file around PAR-CLIP sites according to the given parameters.', epilog="contact: [email protected]")
parser.add_argument('sites', help='PAR-CLIP file *.table')
parser.add_argument('genome', help='path to genome')
parser.add_argument('fafile', help='output filename')
parser.add_argument('filterGFF', help='set path to GFF if sites should be removed that overlap with the GFF [default = '']', default='')
parser.add_argument('start', help='start index of PAR-CLIP sites [default=0]', type=int, default = 0)
parser.add_argument('stop', help='stop index of PAR-CLIP sites [default=1500]', type=int, default = 1500)
parser.add_argument('width', help='number of nt +/- the crosslink site [default=15]', type=int, default = 15)
parser.add_argument('additionalFilterWidth', help='number of nt that are added to the start/stop indices of the GFF annotations', type=int, default = 20)
parser.add_argument('key', help='set key that is used for PAR-CLIP site ordering [default = \'occ\'], options: [\'occ\', \'m\', \'r\', \'mr\', \'pvalue\']', default='occ')
parser.add_argument('-v','--verbose', dest='verbose', action="store_true", default=False, help='verbose output')
args = parser.parse_args()
yeast = genome.Genome(args.genome, False)
sites = ParclipSiteContainer()
sites.loadFromFile(args.sites)
if args.verbose:
print('#sites : '+str(sites.size()))
if args.filterGFF != '':
anno = gff.GFF(args.filterGFF)
sites = sites.removeSitesLocatedInGFF(anno, args.additionalFilterWidth)
print('#sites after removal: '+str(sites.size()))
sites.sort(args.key)
sites.save2Fasta(yeast, args.fafile, args.start, args.stop, args.width)