def __init__(self,gff_file,filetype='standard',fasta_file=None,remove_noncoding=False,limit=None,author=None):
if filetype not in self._filetypes:
raise TypeError()
self.filename = gff_file
if filetype not in self._filetypes:
e = '{0} is not a valid filetype'.format(filetype)
raise TypeError(e)
self.filetype = filetype
self.fasta_file = fasta_file
self.remove_noncoding = remove_noncoding
if limit:
if not isinstance(limit,dict):
e = '{0} is not a valid type for limit'.format(type(limit))
raise TypeError(e)
else:
#reformat limit values into lists if they are not already lists
self.limit = {}
for key,value in limit.iteritems():
if isinstance(value,(list,tuple)):
self.limit[key] = value
elif isinstance(value,basestring):
self.limit[key] = [value]
#self.limit = {key:[value] for key,value in limit.iteritems() if isinstance(value,basestring) or if isinstance(value,(list,tuple))}
else:
self.limit = limit
self.gff = Gff(filename=gff_file)
self.author = author
def __init__(self,gff_file,filetype='standard',fasta_file=None,remove_noncoding=False,limit=None,author=None):
if filetype not in self._filetypes:
raise TypeError()
self.filename = gff_file
if filetype not in self._filetypes:
e = '{0} is not a valid filetype'.format(filetype)
raise TypeError(e)
self.filetype = filetype
self.fasta_file = fasta_file
self.remove_noncoding = remove_noncoding
if limit:
if not isinstance(limit,dict):
e = '{0} is not a valid type for limit'.format(type(limit))
raise TypeError(e)
else:
#reformat limit values into lists if they are not already lists
self.limit = {}
for key,value in limit.iteritems():
if isinstance(value,(list,tuple)):
self.limit[key] = value
elif isinstance(value,basestring):
self.limit[key] = [value]
#self.limit = {key:[value] for key,value in limit.iteritems() if isinstance(value,basestring) or if isinstance(value,(list,tuple))}
else:
self.limit = limit
self.gff = Gff(filename=gff_file)
self.author = author
if keepinfo:
writer.cnames.append('ORIGINAL')
def getNameFromAttrs(attrs):
if attr2name:
return attr2name(**attrs)
for key in sorted(attrs.keys()):
if key in writer.cnames:
continue
if 'id' in key.lower():
return attrs[key]
if 'name' in key.lower():
return attrs[key]
return attrs[key]
gff = Gff(infile)
for record in gff:
r = TsvRecord()
r.CHR = record['seqid']
r.START = record['start']
r.END = record['end']
r.SCORE = record['score']
r.STRAND = record['strand']
attrs = record['attributes']
attrs.update(dict(
CHR = r.CHR,
START = r.START,
END = r.END,
SCORE = r.SCORE,
STRAND = r.STRAND
))
class Parser(object):
"""
Parser object gff3 formatted annotation files.
The parse() method return the processed Gff object.
Has several methods for known incorrectly formatted files: _ratt() and _manual(), which are called by specifying filetype
"""
_filetypes = ('standard','ratt','manual','interproscan')
def __init__(self,gff_file,filetype='standard',fasta_file=None,remove_noncoding=False,limit=None,author=None):
if filetype not in self._filetypes:
raise TypeError()
self.filename = gff_file
if filetype not in self._filetypes:
e = '{0} is not a valid filetype'.format(filetype)
raise TypeError(e)
self.filetype = filetype
self.fasta_file = fasta_file
self.remove_noncoding = remove_noncoding
if limit:
if not isinstance(limit,dict):
e = '{0} is not a valid type for limit'.format(type(limit))
raise TypeError(e)
else:
#reformat limit values into lists if they are not already lists
self.limit = {}
for key,value in limit.iteritems():
if isinstance(value,(list,tuple)):
self.limit[key] = value
elif isinstance(value,basestring):
self.limit[key] = [value]
#self.limit = {key:[value] for key,value in limit.iteritems() if isinstance(value,basestring) or if isinstance(value,(list,tuple))}
else:
self.limit = limit
self.gff = Gff(filename=gff_file)
self.author = author
#self.parsed = False
def _readlines(self):
"""
Generator that yields GffSubPart objects formatted according to filetype.
Limit is a dictionary that can limit the lines parsed based on some attributes like
"""
with open(self.filename,'rU') as fh:
for line in fh:
if not line.strip():
continue
if line[0] == '#':
if 'FASTA' in line:
return
continue
if self.filetype == 'interproscan':
line = re.sub('; ',': ',line)
parts = line.strip().split('\t')
if self.limit != None:
#Series of list comprehensions to check if any of the relevant attributes of sub are in the limit dictionary
#An empty list is falsy, so if the relevant value is not in the limit dictionary, continue to the next line
if not [x for x in self.limit.get('seqid',[parts[0]]) if x == parts[0]]:
continue
elif not [x for x in self.limit.get('source',[parts[1]]) if x == parts[1]]:
continue
elif not [x for x in self.limit.get('featuretype',[parts[2]]) if x == parts[2]]:
continue
elif not [x for x in self.limit.get('strand',[parts[6]]) if x == parts[6]]:
continue
sub = GffSubPart(*parts,filetype=self.filetype)
yield sub
def _ratt(self):
name_index_map = {}
name_index_remove = set()
child_map = {}
for transcript in self.gff.getitems(featuretype='mRNA'):
new_transcript_ID = transcript.attributes['locus_tag'][0]
if transcript.ID == new_transcript_ID:
continue
child_map.setdefault(new_transcript_ID,[])
name_index_map[new_transcript_ID] = transcript._key
name_index_remove.add(transcript.ID)
sub_counter = {}
for sub in self.gff.get_children(transcript):
if sub == transcript:
continue
sub_counter.setdefault(sub.featuretype,0)
sub_counter[sub.featuretype] += 1
sub.parents = [new_transcript_ID]
new_sub_ID = '{0}.{1}{2}'.format(new_transcript_ID,sub.featuretype,sub_counter[sub.featuretype])
child_map.setdefault(new_transcript_ID,[]).append(new_sub_ID)
name_index_map[new_sub_ID] = sub._key
name_index_remove.add(sub.ID)
sub.ID = new_sub_ID
sub.source = 'ratt'
transcript.ID = new_transcript_ID
transcript.source = 'ratt'
transcript.attributes = {key:value for key,value in transcript.attributes.iteritems() if key in ('ID','Parent')}
gene_ID = transcript.ID.split('.')[0]
if not gene_ID in self.gff.name_index:
gene_parts = transcript.gff_fields
gene = GffSubPart(*gene_parts)
gene.featuretype = 'gene'
gene.ID = gene_ID
gene.children = [transcript.ID]
self.gff.update(gene)
else:
for gene in self.gff[gene_ID]:
gene.children.append(transcript.ID)
transcript.parents = [gene_ID]
for new_name,_key in name_index_map.iteritems():
self.gff.name_index[new_name] = [_key]
for old_name in name_index_remove:
self.gff.name_index.pop(old_name)
for parent_id in child_map:
for parent in self.gff[parent_id]:
parent.children = child_map[parent_id]
def _manual(self):
ann = {}
for sub in self.gff.getitems(featuretype='CDS'):
if sub.attributes.get('Name',False):
ann.setdefault(sub.ID,[]).append(sub)
for gene,cds_list in ann.iteritems():
self._assert_equal(cds_list)
template = cds_list[0]
exon_counter = 0
#GffSubPart elements
seqid = template.seqid
source = template.source
start = min(s.start for s in cds_list)
end = max(s.end for s in cds_list)
score = template.score
strand = template.strand
phase = template.phase
#GffSubPart attributes
gene_ID = uuid.uuid4()
mrna_ID = '{0}.1'.format(gene_ID)
name = template.attributes.get('Name',False)[0]
author = self.author#template.attributes.get('Created by',False)[0]
gene_attributes = 'ID={0};Name={1};Created by={2}'.format(gene_ID,name,author)
mrna_attributes = 'ID={0};Parent={1};Name={2};Created by={3}'.format(mrna_ID,gene_ID,name,author)
#make the gene GffSubPart
gene = GffSubPart(seqid,source,'gene',start,end,score,strand,phase,gene_attributes)
#make the mRNA GffSubPart
mrna = GffSubPart(seqid,source,'mRNA',start,end,score,strand,phase,mrna_attributes)
self.gff.update(gene)
self.gff.update(mrna)
if gene.strand == '+':
reverse = True
else:
reverse = False
for cds in sorted(cds_list,key=lambda x:x.get_start(),reverse=reverse):
exon_counter += 1
old_ID = cds.ID
cds.ID = '{0}.CDS{1}'.format(mrna_ID,exon_counter)
cds.attributes['ID'] = ['{0}.CDS{1}'.format(mrna_ID,exon_counter)]
#print self.gff.name_index[old_ID]
self.gff.name_index.setdefault(cds.ID,set()).add(cds._key)
self.gff.name_index[old_ID] = {x for x in self.gff.name_index[old_ID] if x != cds._key}
if len(self.gff.name_index[old_ID]) == 0:
self.gff.name_index.pop(old_ID)
cds.parents = [mrna_ID]
cds.attributes['Parent'] = [mrna_ID]
cds.attributes.pop('modified by',[])
cds.attributes.pop('created by',[])
cds.attributes['Created by'] = [self.author]
def _interproscan(self):
type_dict = {}
for feature in self.gff:
for attribute,values in feature.attributes.iteritems():
feature.attributes[attribute] = [value.strip('"') for value in values]
if feature.featuretype == 'protein_match':
#set parent attribute
parent = feature.attributes['Target'][0]
parent = parent.split()[0]
feature.parents = [parent]
#determine new ID based on parent name and source, this should result in unique IDs, opposed to what comes out of IPS
source = feature.source
type_dict.setdefault(parent,{}).setdefault(source,0)
type_dict[parent][source] += 1
type_count = type_dict[parent][source]
ID = '{0}.{1}.{2}'.format(parent,source,type_count)
self.gff.name_index[ID] = self.gff.name_index.pop(feature.ID)
feature.ID = ID
@staticmethod
def _assert_equal(subslist):
first = subslist[0]
assert all(x.seqid == first.seqid for x in subslist)
assert all(x.source == first.source for x in subslist)
assert all(x.strand == first.strand for x in subslist)
try:
assert all(x.attributes['Name'] == first.attributes['Name'] for x in subslist),[x.attributes['Name'] for x in subslist]
except KeyError as e:
print first
raise e
except AssertionError as e:
raise e
try:
pass
#assert all(x.attributes['Created by'] == first.attributes['Created by'] for x in subslist)
except KeyError as e:
print first
raise e
def _remove_noncoding(self):
if not self.fasta_file:
e = 'Can not remove non coding annotations without sequence data'
raise NotImplementedError(e)
remove = []
for feature in self.gff.getitems(featuretype='mRNA'):
try:
pep = feature.pep
except TranslateError:
remove.append(feature.ID)
continue
if pep[0] != 'M':
if feature.seq[0:3] != 'CTG':
remove.append(feature.ID)
elif pep[-1] != '*':
remove.append(feature.ID)
elif '*' in pep[1:-1]:
remove.append(feature.ID)
print 'Remove {0} genes because they do not encode a protein'.format(len(remove))
self.gff.remove(remove)
def _get_nested_ratt_ids(self,ID):
IDs = set()
for lower_ID in self._get_lower_ratt_ids(ID):
IDs.add(lower_ID)
yield lower_ID
for higher_ID in self._get_higher_ratt_ids(ID):
if higher_ID not in IDs:
yield higher_ID
def _get_lower_ratt_ids(self,ID):
yield ID
if ID[-4] != '.':
return
for sub_ID in self._get_lower_ratt_ids(ID[:-2]):
yield sub_ID
def _get_higher_ratt_ids(self,ID):
yield ID
counter = int(ID[-1])
new_ID = '{0}.{1}'.format(ID,counter + 1)
if new_ID not in self.gff:
return
for super_ID in self._get_higher_ratt_ids(new_ID):
yield super_ID
def _remove_fragmented_ratt(self):
remove = set()
ID_map = {}
for transcript in self.gff.getitems(featuretype='mRNA'):
if transcript.ID in remove:
continue
if transcript.ID[-4] == '.':
#get all fragments
fragments = (ff for f in self._get_nested_ratt_ids(transcript.ID) for ff in self.gff[f])
#sort fragments by CDS length
sorted_fragments = sorted(fragments,key = lambda x:len(x.seq))
#remove the longest fragment from the list so it doesnt get removed
longest = sorted_fragments.pop()
#remove all the other fragments
remove |= {f.ID for f in sorted_fragments}#set(sorted_fragments)
#make sure longest fragment keeps the original name
transcript_ID = [ID for ID in self._get_nested_ratt_ids(longest.ID) if ID[-4] != '.'][0]
ID_map[longest.ID] = transcript_ID
#set ID and Parent attribute of nested CDS features
sub_counter = {}
for sub in self.gff.get_children(longest):
if sub == longest:
continue
sub.attributes['Parent'] = [transcript_ID]
sub_counter.setdefault(sub.featuretype,0)
sub_counter[sub.featuretype] += 1
sub_ID = '{0}.{1}{2}'.format(transcript_ID,sub.featuretype,sub_counter[sub.featuretype])
ID_map[sub.ID] = sub_ID
self.gff.remove(remove)
for old_ID,new_ID in ID_map.iteritems():
assert old_ID not in remove,(old_ID,new_ID)
for feature in self.gff[old_ID]:
feature.ID = new_ID
def parse(self):
self.remove_fragmented_ratt = False
for subpart in self._readlines():
#if self.filetype == 'interproscan':
# if subpart.featuretype == 'protein_match':
# parent = subpart.attributes['Target'][0]
# parent = parent.split()[0]
# subpart.parents = [parent]
self.gff.update(subpart)
if self.filetype == 'ratt':
self.remove_fragmented_ratt = True
self.gff.set_children()
self._ratt()
elif self.filetype == 'manual':
self._manual()
self.gff.set_children()
elif self.filetype == 'interproscan':
self._interproscan()
self.gff.set_children()
else:
pass
if self.filetype != 'ratt':
self.gff.set_children()
if self.fasta_file:
self.gff.add_fasta(self.fasta_file)
if self.remove_fragmented_ratt:
self._remove_fragmented_ratt()
if self.remove_noncoding:
self._remove_noncoding()
self.parsed = True
return self.gff
:param t:
:return:
"""
print('C {}\t{}\t{}'.format(t['transcript_id'], t['begin'], t['end']))
if t['transcript_id'] not in merge['names']:
if merge['strand'] != t['strand']:
print('warning different strands {}'.format(t['transcript_id']))
merge['begin'] = min(merge['begin'], t['begin'])
merge['end'] = min(merge['end'], t['end'])
merge['member'].append(t)
merge['names'].append(t['transcript_id'])
if __name__ == '__main__':
sgff = Gff(file="stranded.merged.stringtie.gff")
transcript_n = sgff.read_feature('transcript')
print('{} stranded features read'.format(transcript_n))
sgff.replace_columns_re(['sequence'], 'lcl\|', r'')
# sgff.attr_sep = '='
# transcript_n = sgff.read_feature('mRNA')
query = r'(maker|augustus|masked|processed|gene|trnascan)(-|_)'
sgff.replace_columns_re(['gene_id', 'transcript_id'], query, r'')
sgff.replace_columns_re(['gene_id', 'transcript_id'], r'-tRNA-', r'.tRNA')
sgff.position_to_int()
# sgff.rename_key('ID', 'transcript_id')
sbundle = make_bundle(sgff)
ugff = Gff(file="unstranded.merged.stringtie.gff")
transcript_n = ugff.read_feature('transcript')
outfile = {{o.outfile | quote}}
notfound = {{args.notfound | quote}}
genecol = {{args.genecol or 0 | repr}}
inopts = {{args.inopts | repr}}
refgene = {{args.refgene | quote}}
if not path.isfile(refgene):
raise OSError('Refgene file does not exists: {}'.format(refgene))
# get genes
genes = TsvReader(infile, **inopts).dump(genecol)
genes = dict(zip(genes, [False] * len(genes)))
writer = TsvWriter(outfile)
writer.cnames = ['CHR', 'START', 'END', 'NAME', 'SCORE', 'STRAND']
gff = Gff(refgene)
for g in gff:
attrs = g['attributes']
if attrs['gene_id'] not in genes:
continue
r = TsvRecord()
r.CHR = g['seqid']
r.START = g['start']
r.END = g['end']
r.SCORE = g['score']
r.STRAND = g['strand']
r.NAME = attrs['gene_id']
writer.write(r)
writer.close()
for g, v in genes.items():
class Parser(object):
"""
Parser object gff3 formatted annotation files.
The parse() method return the processed Gff object.
Has several methods for known incorrectly formatted files: _ratt() and _manual(), which are called by specifying filetype
"""
_filetypes = ('standard','ratt','manual','interproscan')
def __init__(self,gff_file,filetype='standard',fasta_file=None,remove_noncoding=False,limit=None,author=None):
if filetype not in self._filetypes:
raise TypeError()
self.filename = gff_file
if filetype not in self._filetypes:
e = '{0} is not a valid filetype'.format(filetype)
raise TypeError(e)
self.filetype = filetype
self.fasta_file = fasta_file
self.remove_noncoding = remove_noncoding
if limit:
if not isinstance(limit,dict):
e = '{0} is not a valid type for limit'.format(type(limit))
raise TypeError(e)
else:
#reformat limit values into lists if they are not already lists
self.limit = {}
for key,value in limit.iteritems():
if isinstance(value,(list,tuple)):
self.limit[key] = value
elif isinstance(value,basestring):
self.limit[key] = [value]
#self.limit = {key:[value] for key,value in limit.iteritems() if isinstance(value,basestring) or if isinstance(value,(list,tuple))}
else:
self.limit = limit
self.gff = Gff(filename=gff_file)
self.author = author
#self.parsed = False
def _readlines(self):
"""
Generator that yields GffSubPart objects formatted according to filetype.
Limit is a dictionary that can limit the lines parsed based on some attributes like
"""
with open(self.filename,'rU') as fh:
for line in fh:
if not line.strip():
continue
if line[0] == '#':
if 'FASTA' in line:
return
continue
if self.filetype == 'interproscan':
line = re.sub('; ',': ',line)
parts = line.strip().split('\t')
if self.limit != None:
#Series of list comprehensions to check if any of the relevant attributes of sub are in the limit dictionary
#An empty list is falsy, so if the relevant value is not in the limit dictionary, continue to the next line
if not [x for x in self.limit.get('seqid',[parts[0]]) if x == parts[0]]:
continue
elif not [x for x in self.limit.get('source',[parts[1]]) if x == parts[1]]:
continue
elif not [x for x in self.limit.get('featuretype',[parts[2]]) if x == parts[2]]:
continue
elif not [x for x in self.limit.get('strand',[parts[6]]) if x == parts[6]]:
continue
sub = GffSubPart(*parts,filetype=self.filetype)
yield sub
def _ratt(self):
name_index_map = {}
name_index_remove = set()
child_map = {}
for transcript in self.gff.getitems(featuretype='mRNA'):
new_transcript_ID = transcript.attributes['locus_tag'][0]
if transcript.ID == new_transcript_ID:
continue
child_map.setdefault(new_transcript_ID,[])
name_index_map[new_transcript_ID] = transcript._key
name_index_remove.add(transcript.ID)
sub_counter = {}
for sub in self.gff.get_children(transcript):
if sub == transcript:
continue
sub_counter.setdefault(sub.featuretype,0)
sub_counter[sub.featuretype] += 1
sub.parents = [new_transcript_ID]
new_sub_ID = '{0}.{1}{2}'.format(new_transcript_ID,sub.featuretype,sub_counter[sub.featuretype])
child_map.setdefault(new_transcript_ID,[]).append(new_sub_ID)
name_index_map[new_sub_ID] = sub._key
name_index_remove.add(sub.ID)
sub.ID = new_sub_ID
sub.source = 'ratt'
transcript.ID = new_transcript_ID
transcript.source = 'ratt'
transcript.attributes = {key:value for key,value in transcript.attributes.iteritems() if key in ('ID','Parent')}
gene_ID = transcript.ID.split('.')[0]
if not gene_ID in self.gff.name_index:
gene_parts = transcript.gff_fields
gene = GffSubPart(*gene_parts)
gene.featuretype = 'gene'
gene.ID = gene_ID
gene.children = [transcript.ID]
self.gff.update(gene)
else:
for gene in self.gff[gene_ID]:
gene.children.append(transcript.ID)
transcript.parents = [gene_ID]
for new_name,_key in name_index_map.iteritems():
self.gff.name_index[new_name] = [_key]
for old_name in name_index_remove:
self.gff.name_index.pop(old_name)
for parent_id in child_map:
for parent in self.gff[parent_id]:
parent.children = child_map[parent_id]
def _manual(self):
ann = {}
for sub in self.gff.getitems(featuretype='CDS'):
if sub.attributes.get('Name',False):
ann.setdefault(sub.ID,[]).append(sub)
for gene,cds_list in ann.iteritems():
self._assert_equal(cds_list)
template = cds_list[0]
exon_counter = 0
#GffSubPart elements
seqid = template.seqid
source = template.source
start = min(s.start for s in cds_list)
end = max(s.end for s in cds_list)
score = template.score
strand = template.strand
phase = template.phase
#GffSubPart attributes
gene_ID = uuid.uuid4()
mrna_ID = '{0}.1'.format(gene_ID)
name = template.attributes.get('Name',False)[0]
author = self.author#template.attributes.get('Created by',False)[0]
gene_attributes = 'ID={0};Name={1};Created by={2}'.format(gene_ID,name,author)
mrna_attributes = 'ID={0};Parent={1};Name={2};Created by={3}'.format(mrna_ID,gene_ID,name,author)
#make the gene GffSubPart
gene = GffSubPart(seqid,source,'gene',start,end,score,strand,phase,gene_attributes)
#make the mRNA GffSubPart
mrna = GffSubPart(seqid,source,'mRNA',start,end,score,strand,phase,mrna_attributes)
self.gff.update(gene)
self.gff.update(mrna)
if gene.strand == '+':
reverse = True
else:
reverse = False
for cds in sorted(cds_list,key=lambda x:x.get_start(),reverse=reverse):
exon_counter += 1
old_ID = cds.ID
cds.ID = '{0}.CDS{1}'.format(mrna_ID,exon_counter)
cds.attributes['ID'] = ['{0}.CDS{1}'.format(mrna_ID,exon_counter)]
#print self.gff.name_index[old_ID]
self.gff.name_index.setdefault(cds.ID,set()).add(cds._key)
self.gff.name_index[old_ID] = {x for x in self.gff.name_index[old_ID] if x != cds._key}
if len(self.gff.name_index[old_ID]) == 0:
self.gff.name_index.pop(old_ID)
cds.parents = [mrna_ID]
cds.attributes['Parent'] = [mrna_ID]
cds.attributes.pop('modified by',[])
cds.attributes.pop('created by',[])
cds.attributes['Created by'] = [self.author]
def _interproscan(self):
type_dict = {}
for feature in self.gff:
for attribute,values in feature.attributes.iteritems():
feature.attributes[attribute] = [value.strip('"') for value in values]
if feature.featuretype == 'protein_match':
#set parent attribute
parent = feature.attributes['Target'][0]
parent = parent.split()[0]
feature.parents = [parent]
#determine new ID based on parent name and source, this should result in unique IDs, opposed to what comes out of IPS
source = feature.source
type_dict.setdefault(parent,{}).setdefault(source,0)
type_dict[parent][source] += 1
type_count = type_dict[parent][source]
ID = '{0}.{1}.{2}'.format(parent,source,type_count)
self.gff.name_index[ID] = self.gff.name_index.pop(feature.ID)
feature.ID = ID
@staticmethod
def _assert_equal(subslist):
first = subslist[0]
assert all(x.seqid == first.seqid for x in subslist)
assert all(x.source == first.source for x in subslist)
assert all(x.strand == first.strand for x in subslist)
try:
assert all(x.attributes['Name'] == first.attributes['Name'] for x in subslist),[x.attributes['Name'] for x in subslist]
except KeyError as e:
print first
raise e
except AssertionError as e:
raise e
try:
pass
#assert all(x.attributes['Created by'] == first.attributes['Created by'] for x in subslist)
except KeyError as e:
print first
raise e
def _remove_noncoding(self):
if not self.fasta_file:
e = 'Can not remove non coding annotations without sequence data'
raise NotImplementedError(e)
remove = []
for feature in self.gff.getitems(featuretype='mRNA'):
try:
pep = feature.pep
except TranslateError:
remove.append(feature.ID)
continue
if pep[0] != 'M':
if feature.seq[0:3] != 'CTG':
remove.append(feature.ID)
elif pep[-1] != '*':
remove.append(feature.ID)
elif '*' in pep[1:-1]:
remove.append(feature.ID)
print 'Remove {0} genes because they do not encode a protein'.format(len(remove))
self.gff.remove(remove)
def _get_nested_ratt_ids(self,ID):
IDs = set()
for lower_ID in self._get_lower_ratt_ids(ID):
IDs.add(lower_ID)
yield lower_ID
for higher_ID in self._get_higher_ratt_ids(ID):
if higher_ID not in IDs:
yield higher_ID
def _get_lower_ratt_ids(self,ID):
yield ID
if ID[-4] != '.':
return
for sub_ID in self._get_lower_ratt_ids(ID[:-2]):
yield sub_ID
def _get_higher_ratt_ids(self,ID):
yield ID
counter = int(ID[-1])
new_ID = '{0}.{1}'.format(ID,counter + 1)
if new_ID not in self.gff:
return
for super_ID in self._get_higher_ratt_ids(new_ID):
yield super_ID
def _remove_fragmented_ratt(self):
remove = set()
ID_map = {}
for transcript in self.gff.getitems(featuretype='mRNA'):
if transcript.ID in remove:
continue
if transcript.ID[-4] == '.':
#get all fragments
fragments = (ff for f in self._get_nested_ratt_ids(transcript.ID) for ff in self.gff[f])
#sort fragments by CDS length
sorted_fragments = sorted(fragments,key = lambda x:len(x.seq))
#remove the longest fragment from the list so it doesnt get removed
longest = sorted_fragments.pop()
#remove all the other fragments
remove |= {f.ID for f in sorted_fragments}#set(sorted_fragments)
#make sure longest fragment keeps the original name
transcript_ID = [ID for ID in self._get_nested_ratt_ids(longest.ID) if ID[-4] != '.'][0]
ID_map[longest.ID] = transcript_ID
#set ID and Parent attribute of nested CDS features
sub_counter = {}
for sub in self.gff.get_children(longest):
if sub == longest:
continue
sub.attributes['Parent'] = [transcript_ID]
sub_counter.setdefault(sub.featuretype,0)
sub_counter[sub.featuretype] += 1
sub_ID = '{0}.{1}{2}'.format(transcript_ID,sub.featuretype,sub_counter[sub.featuretype])
ID_map[sub.ID] = sub_ID
self.gff.remove(remove)
for old_ID,new_ID in ID_map.iteritems():
assert old_ID not in remove,(old_ID,new_ID)
for feature in self.gff[old_ID]:
feature.ID = new_ID
def parse(self):
self.remove_fragmented_ratt = False
for subpart in self._readlines():
#if self.filetype == 'interproscan':
# if subpart.featuretype == 'protein_match':
# parent = subpart.attributes['Target'][0]
# parent = parent.split()[0]
# subpart.parents = [parent]
self.gff.update(subpart)
if self.filetype == 'ratt':
self.remove_fragmented_ratt = True
self.gff.set_children()
self._ratt()
elif self.filetype == 'manual':
self._manual()
self.gff.set_children()
elif self.filetype == 'interproscan':
self._interproscan()
self.gff.set_children()
else:
pass
if self.filetype != 'ratt':
self.gff.set_children()
if self.fasta_file:
self.gff.add_fasta(self.fasta_file)
if self.remove_fragmented_ratt:
self._remove_fragmented_ratt()
if self.remove_noncoding:
self._remove_noncoding()
self.parsed = True
return self.gff
