def write_new_bed(gene_list, old_bed, missed_genes,out_file):
merge_fh = open(out_file,"wb")
hit_list = [hit for hit,qaccn in missed_genes]
for i,gene in enumerate(old_bed):
if gene["accn"] in hit_list: continue
new_line = Bed.row_string(gene)
merge_fh.write("{0}\n".format(new_line))
for i,new_gene in enumerate(gene_list):
### merge overlapping here
updated_feat = gene_list[new_gene]
if len(updated_feat["locs"]) > 1:
updated_feat = merge_feats(updated_feat)
new_line = Bed.row_string(updated_feat)
merge_fh.write("{0}\n".format(new_line))
def write_new_bed(gene_list, old_bed, missed_genes, out_file):
merge_fh = open(out_file, "wb")
hit_list = [hit for hit, qaccn in missed_genes]
for i, gene in enumerate(old_bed):
if gene["accn"] in hit_list: continue
new_line = Bed.row_string(gene)
merge_fh.write("{0}\n".format(new_line))
for i, new_gene in enumerate(gene_list):
### merge overlapping here
updated_feat = gene_list[new_gene]
if len(updated_feat["locs"]) > 1:
updated_feat = merge_feats(updated_feat)
new_line = Bed.row_string(updated_feat)
merge_fh.write("{0}\n".format(new_line))
def print_bed(flist, old_path):
ipath, ext = op.splitext(old_path)
path = "%s.with_new%s" % (ipath, ext)
print >> sys.stderr, "writing to: %s.with_new%s" % (ipath, ext)
fh = open(path, 'wb')
seen = {}
for item in flist:
# convert the locs to a tuple.
#print >>sys.stderr, item
item = list(item)
item[6] = tuple(item[6])
item = tuple(item)
if item in seen: continue
seen[item] = 1
locs = item[6] # tuple(sorted([item[1], item[2]]))
row = dict(accn=item[3],
start=item[1],
end=item[2],
seqid=item[0],
locs=locs,
score='.',
strand=item[5],
rgb='.',
thickstart='.',
thickend=".")
print >> fh, Bed.row_string(row)
fh.close()
return Bed(path)
def merge(org_bed, missed, merge_file):
"""creates blast.all file and updates everything"""
merge_fh = open(merge_file, "w")
#cds_missed = missed[missed['ftype'] == 'CDS']
#count = org_bed.shape[0] + missed[missed['ftype'] !='CDS'].shape[0]
new_rows = []
seen_accns = {}
# CDS added to existing gene.
for row_missed in missed:
if row_missed['accn'] in seen_accns: continue
try:
org_bed_row = org_bed.accn(row_missed['accn'])
# it's a CDS
except KeyError:
#its a new gene
new_rows.append(row_missed)
seen_accns[row_missed['accn']] = True
continue
locs_interval = Intersecter()
[locs_interval.add_interval(Feature(start,stop)) for start,stop in org_bed_row['locs']]
for missed_start,missed_end in row_missed['locs']:
if len(locs_interval.find(missed_start,missed_end)) > 0:
# print >>sys.stderr, org_bed_row['accn']
locs_intersects = [(l.start,l.stop) for l in locs_interval.find(missed_start,missed_end)]
[org_bed_row['locs'].remove(locs_intersect) for locs_intersect in locs_intersects]
locs_intersects = set(locs_intersects)
locs_intersects.add((missed_start,missed_end))
locs_start = min([start for start,end in locs_intersects])
locs_end = max([end for start,end in locs_intersects])
org_bed_row['locs'] = org_bed_row['locs'] + [(locs_start,locs_end)]
row_missed['locs'].remove((missed_start,missed_end))
org_bed_row['locs'] = org_bed_row['locs'] + row_missed['locs']
#print >>sys.stderr, "{0},{1}".format(row_missed['accn'], locs)
org_bed_row['locs'].sort()
org_bed_row['start'] = min(min([start for start,end in org_bed_row['locs']]), org_bed_row['start'])
org_bed_row['end'] = max(max([end for start,end in org_bed_row['locs']]), org_bed_row['end'])
new_rows.append(org_bed_row)
seen_accns[org_bed_row['accn']] =True
for org_bed_rw in org_bed:
if org_bed_rw['accn'] not in seen_accns:
new_rows.append(org_bed_rw)
seen_accns[org_bed_rw['accn']] =True
def row_cmp(a,b):
return cmp(a['seqid'], b['seqid']) or cmp(a['start'], b['start'])
new_rows.sort(cmp=row_cmp)
#print >>merge_fh, "\t".join(Bed.names)
for i, row in enumerate(new_rows):
print >>merge_fh, Bed.row_string(row)
def merge_same_hits(missed, fh_match, org_bed):
""" groups genes that hit more then once """
d = {}
handle = open(fh_match)
matches = handle.read()
org_bed_path = org_bed.path
path = org_bed_path.split('/')
dirc = '/'.join(path[:-1])
org = path[-1]
fh = open('{0}/missed_from_{1}'.format(dirc,org), "wb")
for match in matches.split('\n')[:-1]:
qaccn,saccn = match.split('\t')
#create dictionary
try:
seqid = missed.accn(qaccn)['seqid']
haccn = missed.accn(qaccn)
except KeyError: continue
#if near_gene(haccn,org_bed)==True: continue
if (seqid,saccn) not in d.keys():
#append whole dict to keys
d[(seqid,saccn)]= missed.accn(qaccn)
else:
#else add locs to exsting one
gene_start = min(d[(seqid,saccn)]['locs'])[0]
gene_end = max(d[(seqid,saccn)]['locs'])[1]
missed_end = missed.accn(qaccn)['locs'][0][1]
missed_start = missed.accn(qaccn)['locs'][0][0]
if missed_end < gene_start:
# if no intervening genes and they are close together...
intervening_genes = get_intervening_genes(missed_end,gene_start,seqid, org_bed, d[(seqid,saccn)]['accn'])
if intervening_genes is False:
d[(seqid,saccn)]['locs'] = d[(seqid,saccn)]['locs'] + missed.accn(qaccn)['locs']
d[(seqid,saccn)]['start'] = missed_start
if 'Os' in qaccn:
d[seqid,saccn]['accn'] = qaccn
else:
d[(seqid,qaccn)] = missed.accn(qaccn)
elif gene_end < missed_start:
intervening_genes = get_intervening_genes(gene_end,missed_start,seqid, org_bed,d[(seqid,saccn)]["accn"])
if intervening_genes is False:
d[(seqid,saccn)]['locs'] = d[(seqid,saccn)]['locs'] + missed.accn(qaccn)['locs']
d[(seqid,saccn)]['end'] = missed_end
if 'Os' in qaccn:
d[seqid,saccn]['accn'] = qaccn
else:
d[(seqid,qaccn)]= missed.accn(qaccn)
else:
d[(seqid,saccn)]['locs'] = d[(seqid,saccn)]['locs'] + missed.accn(qaccn)['locs']
for key in d.keys():
new_row = d[key]['locs'].sort()
row = d[key]
print >>fh, Bed.row_string(row)
def merge_flat(new_name, aflat, bflat):
"""take 2 flat files and return a new one that is the union of the 2
existing"""
seen = {}
both = []
for flat in (aflat, bflat):
for row in flat:
key = row['seqid'], row['accn']
if key in seen: continue
seen[key] = True
both.append(row)
both.sort(key=lambda a: (a['seqid'], a['start']))
fh = open(new_name, "w")
#print >>fh, "\t".join(Flat.names)
for b in both:
print >> fh, Bed.row_string(b)
fh.close()
return Bed(fh.name)
def merge_flat(new_name, aflat, bflat):
"""take 2 flat files and return a new one that is the union of the 2
existing"""
seen = {}
both = []
for flat in (aflat, bflat):
for row in flat:
key = row['seqid'], row['accn']
if key in seen: continue
seen[key] = True
both.append(row)
both.sort(key=lambda a: (a['seqid'],a['start']))
fh = open(new_name, "w")
#print >>fh, "\t".join(Flat.names)
for b in both:
print >>fh, Bed.row_string(b)
fh.close()
return Bed(fh.name)
def print_bed(flist, old_path):
ipath, ext = op.splitext(old_path)
path = "%s.with_new%s" % (ipath, ext)
print >>sys.stderr, "writing to: %s.with_new%s" % (ipath, ext)
fh = open(path, 'wb')
seen = {}
for item in flist:
# convert the locs to a tuple.
#print >>sys.stderr, item
item = list(item)
item[6] = tuple(item[6])
item = tuple(item)
if item in seen: continue
seen[item] = 1
locs = item[6] # tuple(sorted([item[1], item[2]]))
row = dict(accn=item[3], start=item[1], end=item[2], seqid=item[0],
locs=locs, score='.', strand=item[5], rgb='.', thickstart='.', thickend=".")
print >>fh, Bed.row_string(row)
fh.close()
return Bed(path)
def merge(org_bed, missed, merge_file):
"""creates blast.all file and updates everything"""
merge_fh = open(merge_file, "w")
#cds_missed = missed[missed['ftype'] == 'CDS']
#count = org_bed.shape[0] + missed[missed['ftype'] !='CDS'].shape[0]
new_rows = []
seen_accns = {}
# CDS added to existing gene.
for row_missed in missed:
if row_missed['accn'] in seen_accns: continue
try:
org_bed_row = org_bed.accn(row_missed['accn'])
# it's a CDS
except KeyError:
#its a new gene
new_rows.append(row_missed)
seen_accns[row_missed['accn']] = True
continue
locs_interval = Intersecter()
[
locs_interval.add_interval(Feature(start, stop))
for start, stop in org_bed_row['locs']
]
for missed_start, missed_end in row_missed['locs']:
if len(locs_interval.find(missed_start, missed_end)) > 0:
# print >>sys.stderr, org_bed_row['accn']
locs_intersects = [
(l.start, l.stop)
for l in locs_interval.find(missed_start, missed_end)
]
[
org_bed_row['locs'].remove(locs_intersect)
for locs_intersect in locs_intersects
]
locs_intersects = set(locs_intersects)
locs_intersects.add((missed_start, missed_end))
locs_start = min([start for start, end in locs_intersects])
locs_end = max([end for start, end in locs_intersects])
org_bed_row['locs'] = org_bed_row['locs'] + [
(locs_start, locs_end)
]
row_missed['locs'].remove((missed_start, missed_end))
org_bed_row['locs'] = org_bed_row['locs'] + row_missed['locs']
#print >>sys.stderr, "{0},{1}".format(row_missed['accn'], locs)
org_bed_row['locs'].sort()
org_bed_row['start'] = min(
min([start for start, end in org_bed_row['locs']]),
org_bed_row['start'])
org_bed_row['end'] = max(
max([end for start, end in org_bed_row['locs']]),
org_bed_row['end'])
new_rows.append(org_bed_row)
seen_accns[org_bed_row['accn']] = True
for org_bed_rw in org_bed:
if org_bed_rw['accn'] not in seen_accns:
new_rows.append(org_bed_rw)
seen_accns[org_bed_rw['accn']] = True
def row_cmp(a, b):
return cmp(a['seqid'], b['seqid']) or cmp(a['start'], b['start'])
new_rows.sort(cmp=row_cmp)
#print >>merge_fh, "\t".join(Bed.names)
for i, row in enumerate(new_rows):
print >> merge_fh, Bed.row_string(row)
def merge_same_hits(missed, fh_match, org_bed):
""" groups genes that hit more then once """
d = {}
handle = open(fh_match)
matches = handle.read()
org_bed_path = org_bed.path
path = org_bed_path.split('/')
dirc = '/'.join(path[:-1])
org = path[-1]
fh = open('{0}/missed_from_{1}'.format(dirc, org), "wb")
for match in matches.split('\n')[:-1]:
qaccn, saccn = match.split('\t')
#create dictionary
try:
seqid = missed.accn(qaccn)['seqid']
haccn = missed.accn(qaccn)
except KeyError:
continue
#if near_gene(haccn,org_bed)==True: continue
if (seqid, saccn) not in d.keys():
#append whole dict to keys
d[(seqid, saccn)] = missed.accn(qaccn)
else:
#else add locs to exsting one
gene_start = min(d[(seqid, saccn)]['locs'])[0]
gene_end = max(d[(seqid, saccn)]['locs'])[1]
missed_end = missed.accn(qaccn)['locs'][0][1]
missed_start = missed.accn(qaccn)['locs'][0][0]
if missed_end < gene_start:
# if no intervening genes and they are close together...
intervening_genes = get_intervening_genes(
missed_end, gene_start, seqid, org_bed,
d[(seqid, saccn)]['accn'])
if intervening_genes is False:
d[(seqid, saccn)]['locs'] = d[
(seqid, saccn)]['locs'] + missed.accn(qaccn)['locs']
d[(seqid, saccn)]['start'] = missed_start
if 'Os' in qaccn:
d[seqid, saccn]['accn'] = qaccn
else:
d[(seqid, qaccn)] = missed.accn(qaccn)
elif gene_end < missed_start:
intervening_genes = get_intervening_genes(
gene_end, missed_start, seqid, org_bed,
d[(seqid, saccn)]["accn"])
if intervening_genes is False:
d[(seqid, saccn)]['locs'] = d[
(seqid, saccn)]['locs'] + missed.accn(qaccn)['locs']
d[(seqid, saccn)]['end'] = missed_end
if 'Os' in qaccn:
d[seqid, saccn]['accn'] = qaccn
else:
d[(seqid, qaccn)] = missed.accn(qaccn)
else:
d[(seqid, saccn)]['locs'] = d[
(seqid, saccn)]['locs'] + missed.accn(qaccn)['locs']
for key in d.keys():
new_row = d[key]['locs'].sort()
row = d[key]
print >> fh, Bed.row_string(row)
def write_bed(gene,merge_fh):
new_line = Bed.row_string(gene)
merge_fh.write("{0}\n".format(new_line))
def write_bed(gene, merge_fh):
new_line = Bed.row_string(gene)
merge_fh.write("{0}\n".format(new_line))
