def __iter__(self):
if self.is_debug:
fname = self.protxml + '.dump'
logging.debug('Dumping protxml reads into ' + fname)
self.debug_file = open(fname, 'w')
self.debug_file.write('{\n')
for event, elem in etree.iterparse(self.protxml, events=('end', 'start-ns')):
if event == 'start-ns':
self.nsmap.update({elem})
if event == 'end':
if elem.tag == parse.fixtag('', 'protein_group', self.nsmap):
group = parse_protein_group(elem, self.nsmap)
yield group
if self.is_debug:
pprint(group, stream=self.debug_file)
self.debug_file.write(',\n')
elem.clear()
elif elem.tag == parse.fixtag('', 'proteinprophet_details', self.nsmap):
self.distribution = parse_protein_probabilities(elem, self.nsmap)
if self.is_debug:
fname = self.protxml + '.distribution.dump'
pprint(self.distribution, open(fname, 'w'))
elem.clear()
if self.is_debug:
self.debug_file.write('}\n')
self.debug_file.close()
def read_protxml(protxml):
nsmap = {}
distribution = {}
protein_groups = {}
for event, elem in etree.iterparse(protxml, events=('end', 'start-ns')):
if event == 'start-ns':
nsmap.update({elem})
if event == 'end':
if elem.tag == parse.fixtag('', 'protein_group', nsmap):
group = parse_protein_group(elem, nsmap)
protein_groups[group['group_number']] = group
elem.clear()
elif elem.tag == parse.fixtag('', 'proteinprophet_details', nsmap):
distribution = parse_protein_probabilities(elem, nsmap)
elem.clear()
return protein_groups, distribution
def parse_scan(scan_elem, nsmap):
scan = parse.parse_attrib(scan_elem)
scan['matches'] = []
tag = lambda tag_id: parse.fixtag('', tag_id, nsmap)
for search_elem in scan_elem.findall(parse.fixtag('', "search_result", nsmap)):
search_hit_elem = search_elem[0]
match = parse.parse_attrib(search_hit_elem)
match['modified_sequence'] = match['peptide']
match['other_seqids'] = []
for alt_protein in search_hit_elem.findall(parse.fixtag('', 'alternative_protein', nsmap)):
match['other_seqids'].append(alt_protein.attrib['protein'])
match['modifications'] = []
for modified_elem in search_hit_elem.findall(parse.fixtag('', 'modification_info', nsmap)):
attr = parse.parse_attrib(modified_elem)
match['modified_sequence'] = attr['modified_peptide']
for modification_elem in modified_elem.findall(parse.fixtag('', 'mod_aminoacid_mass', nsmap)):
attr = parse.parse_attrib(modification_elem)
attr['i'] = attr['position'] - 1
del attr['position']
match['modifications'].append(attr)
for score_elem in search_hit_elem.findall(tag('search_score')):
match.update(parse.parse_name_value(score_elem))
for analysis_elem in search_hit_elem.find(parse.fixtag('', 'analysis_result', nsmap)):
if analysis_elem.tag == parse.fixtag('', 'peptideprophet_result', nsmap):
match.update(parse.parse_attrib(analysis_elem))
for param_elem in analysis_elem[0]:
match.update(parse.parse_name_value(param_elem))
scan['matches'].append(match)
return scan
def parse_peptide_probabilities(elem, nsmap):
# try with error_point
error_points = elem.findall(parse.fixtag('', 'error_point', nsmap))
if len(error_points) == 0:
charge = 0
for charge_elem in elem.findall(parse.fixtag('', 'roc_error_data', nsmap)):
if charge_elem.attrib['charge'] == 'all':
error_points = charge_elem.findall(parse.fixtag('', 'error_point', nsmap))
break
probs = []
for elem in error_points:
attrib = parse.parse_attrib(elem)
probs.append({
'error': attrib['error'],
'prob': attrib['min_prob'],
})
probs.sort(key=lambda d:d['error'])
return probs
def __iter__(self):
if self.is_debug:
fname = self.pepxml + '.dump'
logging.debug('Dumping pepxml reads into ' + fname)
self.debug_file = open(fname, 'w')
self.debug_file.write('[\n')
for event, elem in etree.iterparse(self.pepxml, events=('start', 'end', 'start-ns')):
if event == 'start-ns':
self.nsmap.update({elem})
elif event == 'start':
if elem.tag == parse.fixtag('', 'msms_run_summary', self.nsmap):
fname = elem.attrib['base_name']
self.source_names.append(fname)
self.i_source = len(self.source_names) - 1
elif event == 'end':
if elem.tag == parse.fixtag('', 'spectrum_query', self.nsmap):
scan = parse_scan(elem, self.nsmap)
for match in scan['matches']:
fpe = probability_to_error(self.distribution, match['probability'])
if fpe is None:
print("WTF", match['probability'], self.distribution)
else:
match['fpe'] = fpe
if self.i_source is not None:
scan['source'] = self.source_names[self.i_source]
if self.is_debug:
pprint(scan, stream=self.debug_file)
self.debug_file.write(',\n')
yield scan
elem.clear()
elif elem.tag == parse.fixtag('', 'peptideprophet_summary', self.nsmap):
self.distribution = parse_peptide_probabilities(elem, self.nsmap)
if self.distribution[0]['prob'] < 1.0:
self.distribution.insert(0, {'prob':1.0, 'error':0.0})
if self.distribution[-1]['prob'] > 0.0:
self.distribution.append({'prob':0.0, 'error':1.0})
if self.is_debug:
fname = self.pepxml + '.distribution.dump'
pprint(self.distribution, open(fname, 'w'))
elem.clear()
if self.is_debug:
self.debug_file.write(']\n')
self.debug_file.close()
def parse_protein_probabilities(elem, nsmap):
probs = []
for data_point in elem.findall(parse.fixtag('', 'protein_summary_data_filter', nsmap)):
attrib = parse.parse_attrib(data_point)
probs.append({
'error': attrib['false_positive_error_rate'],
'prob': attrib['min_probability'],
})
probs.sort(key=lambda d:d['error'])
return probs
def parse_protein_group(elem, nsmap):
group = parse.parse_attrib(elem)
group['proteins'] = []
for protein_elem in elem.findall(parse.fixtag('', 'protein', nsmap)):
protein = parse.parse_attrib(protein_elem)
protein['group_number'] = group['group_number']
annotation_elem = protein_elem.find(parse.fixtag('', 'annotation', nsmap))
if annotation_elem is not None:
protein['description'] = annotation_elem.attrib['protein_description']
protein['other_seqids'] = []
for alt_protein in protein_elem.findall(parse.fixtag('', 'indistinguishable_protein', nsmap)):
protein['other_seqids'].append(alt_protein.attrib['protein_name'])
protein['other_seqids'] = protein['other_seqids']
protein['protein_name'] = protein['protein_name']
protein['peptides'] = []
n_unique_peptide = 0
for peptide_elem in protein_elem.findall(parse.fixtag('', 'peptide', nsmap)):
peptide = parse.parse_attrib(peptide_elem)
protein['peptides'].append(peptide)
peptide['modifications'] = []
peptide['modified_sequence'] = peptide['peptide_sequence']
for modified_elem in peptide_elem.findall(parse.fixtag('', 'modification_info', nsmap)):
attr = parse.parse_attrib(modified_elem)
peptide['modified_sequence'] = attr['modified_peptide']
for modification_elem in modified_elem.findall(parse.fixtag('', 'mod_aminoacid_mass', nsmap)):
attr = parse.parse_attrib(modification_elem)
peptide['modifications'].append(attr)
group['proteins'].append(protein)
return group
def iter(self):
for event, elem in etree.iterparse(self.pepxml, events=('start', 'end', 'start-ns')):
if event == 'start-ns':
self.nsmap.update({elem})
elif event == 'start':
if elem.tag == parse.fixtag('', 'msms_run_summary', self.nsmap):
fname = elem.attrib['base_name']
self.source_names.append(fname)
self.i_source = len(self.source_names) - 1
elif event == 'end':
if elem.tag == parse.fixtag('', 'spectrum_query', self.nsmap):
scan = parse_scan(elem, self.nsmap)
if self.i_source is not None:
scan['source'] = self.source_names[self.i_source]
if self.prob_cutoff is None or scan['probability'] >= self.prob_cutoff:
yield scan
elem.clear()
elif elem.tag == parse.fixtag('', 'peptideprophet_summary', self.nsmap):
self.probs = parse_peptide_probabilities(elem, self.nsmap)
if self.prob_cutoff is None and self.error_cutoff is not None:
self.prob_cutoff = error_to_probability(self.probs, self.prob_cutoff)
elem.clear()
def read_pepxml(pepxml):
nsmap = {}
probs = []
scan_sources = []
for event, elem in etree.iterparse(pepxml, events=('start', 'end', 'start-ns')):
if event == 'start-ns':
nsmap.update({elem})
elif event == 'start':
if elem.tag == parse.fixtag('', 'msms_run_summary', nsmap):
scan_source = {
'scans': [],
'filename': elem.attrib['base_name'],
}
scan_sources.append(scan_source)
elif event == 'end':
if elem.tag == parse.fixtag('', 'spectrum_query', nsmap):
scan = parse_scan(elem, nsmap)
scan_source['scans'].append(scan)
elem.clear()
elif elem.tag == parse.fixtag('', 'peptideprophet_summary', nsmap):
probs = parse_peptide_probabilities(elem, nsmap)
elem.clear()
return scan_sources, probs
def parse_protein_group(elem, nsmap):
group = parse.parse_attrib(elem)
group['proteins'] = []
for protein_elem in elem.findall(parse.fixtag('', 'protein', nsmap)):
protein = parse.parse_attrib(protein_elem)
protein['group_number'] = group['group_number']
for parameter_elem in protein_elem.findall(
parse.fixtag('', 'parameter', nsmap)):
key = parameter_elem.attrib['name']
val = parameter_elem.attrib['value']
protein[key] = val
annotation_elem = protein_elem.find(
parse.fixtag('', 'annotation', nsmap))
if annotation_elem is not None:
protein['description'] = annotation_elem.attrib[
'protein_description']
protein['other_seqids'] = []
for alt_protein in protein_elem.findall(
parse.fixtag('', 'indistinguishable_protein', nsmap)):
protein['other_seqids'].append(alt_protein.attrib['protein_name'])
protein['other_seqids'] = protein['other_seqids']
protein['protein_name'] = protein['protein_name']
protein['peptides'] = []
n_unique_peptide = 0
for peptide_elem in protein_elem.findall(
parse.fixtag('', 'peptide', nsmap)):
peptide = parse.parse_attrib(peptide_elem)
protein['peptides'].append(peptide)
peptide['modifications'] = []
peptide['modified_sequence'] = peptide['peptide_sequence']
for modified_elem in peptide_elem.findall(
parse.fixtag('', 'modification_info', nsmap)):
attr = parse.parse_attrib(modified_elem)
peptide['modified_sequence'] = attr['modified_peptide']
for modification_elem in modified_elem.findall(
parse.fixtag('', 'mod_aminoacid_mass', nsmap)):
attr = parse.parse_attrib(modification_elem)
peptide['modifications'].append(attr)
group['proteins'].append(protein)
return group
def search_tag(self, tag):
return parse.fixtag('', tag, self.nsmap)