def update_drugbank_mappings():
"""Update mappings from DrugBank to CHEBI/CHEMBL"""
# Note that for this to work, PyOBO (https://github.com/pyobo/pyobo) has
# to be installed and the DrugBank download
# (https://www.drugbank.ca/releases/latest) put into ~/.obo/drugbank/
# Note that the DrugBank download requires signing up for an account and
# waiting for approval.
import pyobo
drugbank_chembl = pyobo.get_filtered_xrefs('drugbank', 'chembl.compound')
drugbank_chebi = pyobo.get_filtered_xrefs('drugbank', 'chebi')
chebi_drugbank = pyobo.get_filtered_xrefs('chebi', 'drugbank')
drugbank_names = pyobo.get_id_name_mapping('drugbank')
rows = []
for drugbank_id, chembl_id in drugbank_chembl.items():
rows.append([drugbank_id, 'CHEMBL', chembl_id, 'drugbank'])
for drugbank_id, chebi_id in drugbank_chebi.items():
rows.append([drugbank_id, 'CHEBI', chebi_id, 'drugbank'])
for chebi_id, drugbank_id in chebi_drugbank.items():
rows.append([drugbank_id, 'CHEBI', chebi_id, 'chebi'])
for drugbank_id, name in drugbank_names.items():
rows.append([drugbank_id, 'NAME', name, 'drugbank'])
fname = os.path.join(path, 'drugbank_mappings.tsv')
header = ['DRUGBANK_ID', 'NAMESPACE', 'ID', 'SOURCE']
rows = [header] + sorted(rows)
write_unicode_csv(fname, rows, delimiter='\t')
def get_gene_associations_df(identifier: str, *, rows: Optional[int] = None) -> pd.DataFrame:
"""Get gene associations for the given GO identifier as a dataframe.
- filtered for human onlay
- filtered for proteins only
- add HGNC identifier and entrez identifier
"""
associations = get_gene_associations(identifier, rows=rows)
df = pd.DataFrame(
[
(
e['subject']['id'],
e['subject']['label'],
e['subject']['taxon']['id'][len('NCBITaxon:'):],
e['object']['id'],
e['object']['label'],
e['negated'],
# e['relation']['category'],
# e['relation']['id'],
# e['relation']['inverse'],
# e['relation']['label'],
# e['subject_extensions'],
)
for e in associations
],
columns=[
'source_id',
'source_name',
'taxonomy_id',
'target_id',
'target_label',
'negated',
# 'relation_category',
# 'relation_id',
# 'relation_inverse',
# 'relation_label',
# 'subject_extensions',
],
)
df = df[df['taxonomy_id'] == '9606']
df = df[df['source_id'].str.startswith('UniProtKB:')]
df['uniprot_id'] = df['source_id'].map(lambda s: s[len('UniProtKB:'):])
del df['source_id']
del df['taxonomy_id']
df['hgnc_id'] = df['uniprot_id'].map(get_hgnc_id)
df = df[df['hgnc_id'].notna()]
df['ncbigene_id'] = df['hgnc_id'].map(pyobo.get_filtered_xrefs('hgnc', 'ncbigene').__getitem__)
df['target_id'] = df['target_id'].map(lambda s: s[len('GO:'):])
return df
def test_get_target_xrefs(self):
"""Test getting xrefs."""
kegg_xrefs = get_filtered_xrefs('chebi', 'kegg', url=TEST_CHEBI_OBO_PATH, local=True)
print(kegg_xrefs)
for key, value in kegg_xrefs.items():
self.assertFalse(key.startswith('CHEBI:'))
self.assertFalse(key.startswith('CHEBI'))
self.assertFalse(key.startswith('chebi:'))
self.assertFalse(key.startswith('chebi'))
self.assertFalse(value.startswith('KEGG:'))
self.assertFalse(value.startswith('KEGG'))
self.assertFalse(value.startswith('kegg:'))
self.assertFalse(value.startswith('kegg'))
self.assertIsInstance(kegg_xrefs, dict)
def test_get_target_xrefs(self):
"""Test getting xrefs."""
with chebi_patch:
kegg_xrefs = get_filtered_xrefs('chebi', 'kegg')
for key, value in kegg_xrefs.items():
self.assertFalse(key.startswith('CHEBI:'))
self.assertFalse(key.startswith('CHEBI'))
self.assertFalse(key.startswith('chebi:'))
self.assertFalse(key.startswith('chebi'))
self.assertFalse(value.startswith('KEGG:'))
self.assertFalse(value.startswith('KEGG'))
self.assertFalse(value.startswith('kegg:'))
self.assertFalse(value.startswith('kegg'))
self.assertIsInstance(kegg_xrefs, dict)
def test_get_target_xrefs(self):
"""Test getting xrefs."""
with chebi_patch:
kegg_xrefs = get_filtered_xrefs("chebi", "kegg")
for key, value in kegg_xrefs.items():
self.assertFalse(key.startswith("CHEBI:"))
self.assertFalse(key.startswith("CHEBI"))
self.assertFalse(key.startswith("chebi:"))
self.assertFalse(key.startswith("chebi"))
self.assertFalse(value.startswith("KEGG:"))
self.assertFalse(value.startswith("KEGG"))
self.assertFalse(value.startswith("kegg:"))
self.assertFalse(value.startswith("kegg"))
self.assertIsInstance(kegg_xrefs, dict)
def mutual_mapping_graph(
prefixes: Iterable[str],
skip_sources: Optional[Iterable[str]] = None,
skip_targets: Optional[Iterable[str]] = None,
) -> nx.Graph:
"""Get the undirected mapping graph between the given prefixes.
:param prefixes: A list of prefixes to use with :func:`pyobo.get_filtered_xrefs` to get xrefs.
:param skip_sources: An optional list of prefixes to skip as the source for xrefs
:param skip_targets: An optional list of prefixes to skip as the target for xrefs
:return: The undirected mapping graph containing mappings between entries in the given namespaces.
"""
prefixes = sorted(prefixes)
skip_sources = set() if skip_sources is None else set(skip_sources)
skip_targets = set() if skip_targets is None else set(skip_targets)
graph = nx.Graph()
for source, target in itt.product(prefixes, repeat=2):
if source == target or source in skip_sources or target in skip_targets:
continue
for source_id, target_id in pyobo.get_filtered_xrefs(source,
target).items():
graph.add_edge((source, source_id), (target, target_id))
return graph
def populate(self, paths: Optional[Mapping[str, str]] = None):
"""Populate the database.
:param paths: mapping from tax identifiers to paths to GMT files
"""
if not paths:
logger.info('No paths given.')
paths = {info.taxonomy_id: info.path for info in infos.values()}
logger.info(f'Using default paths at {paths}.')
elif not isinstance(paths, dict):
raise TypeError('Invalid type for paths. Shoudl be dict.')
pathways = [
pathway
for taxonomy_id, path in paths.items()
for pathway in parse_wikipathways_gmt(path)
]
versions = {
version
for _identifier, version, _revision, _name, _species_name, _entries in pathways
}
if len(versions) != 1:
raise ValueError('got multiple versions')
version = list(versions)[0]
taxonomy_name_to_id = get_name_id_mapping('ncbitaxon')
species_names = {
SPECIES_REMAPPING.get(species_name, species_name)
for _identifier, _version, _revision, _name, species_name, _entries in pathways
}
species_name_to_species = {}
for species_name in tqdm(species_names, desc=f'v{version} serializing species'):
taxonomy_id = taxonomy_name_to_id[species_name]
species = species_name_to_species[species_name] = Species(taxonomy_id=taxonomy_id, name=species_name)
self.session.add(species)
hgnc_id_to_entrez_id = get_filtered_xrefs('hgnc', 'ncbigene')
if not hgnc_id_to_entrez_id:
raise ValueError('Mappings from hgnc to ncbigene couldnt be loaded')
entrez_id_to_hgnc_id = {v: k for k, v in hgnc_id_to_entrez_id.items()}
hgnc_id_to_name = get_id_name_mapping('hgnc')
missing_entrez_ids = set()
entrez_ids = {
entrez_id
for _identifier, _version, _revision, _name, _species, entrez_ids in pathways
for entrez_id in entrez_ids
}
entrez_id_protein = {}
for entrez_id in tqdm(entrez_ids, desc=f'v{version} serializing proteins'):
hgnc_id = entrez_id_to_hgnc_id.get(entrez_id)
if hgnc_id:
hgnc_symbol = hgnc_id_to_name[hgnc_id]
else:
hgnc_symbol = None
if not hgnc_symbol:
logging.debug(f"ncbigene:{entrez_id} has no HGNC identifier")
missing_entrez_ids.add(entrez_id)
entrez_id_protein[entrez_id] = protein = self.get_or_create_protein(
entrez_id=entrez_id,
hgnc_symbol=hgnc_symbol,
hgnc_id=hgnc_id,
)
self.session.add(protein)
logger.info(f'Proteins: {len(entrez_id_protein)}')
logger.info(f"Proteins w/o HGNC mapping: {len(missing_entrez_ids)}")
for (
wikipathways_id, _version, revision,
pathway_name, species_name, entrez_ids,
) in tqdm(pathways, desc=f'v{version} serializing pathways'):
proteins = [
entrez_id_protein[entrez_id]
for entrez_id in entrez_ids
]
pathway = self.get_or_create_pathway(
identifier=wikipathways_id,
name=pathway_name.strip(),
revision=revision,
species=species_name_to_species[SPECIES_REMAPPING.get(species_name, species_name)],
proteins=proteins,
)
self.session.add(pathway)
self.session.commit()
def _map_hgnc_to_entrez(hgnc_id):
return get_filtered_xrefs('hgnc', 'ncbigene').get(hgnc_id)
import pybel
import pybel.dsl
from pybel import BELGraph
from ..compath import CompathManager, CompathPathwayMixin, CompathProteinMixin
from ..utils import get_data_dir
logger = logging.getLogger(__name__)
MODULE_NAME = 'pid'
DIRECTORY = get_data_dir(MODULE_NAME)
URL = 'https://github.com/NCIP/pathway-interaction-database/raw/master/download/NCI-Pathway-Info.xlsx'
chebi_id_to_name = get_id_name_mapping('chebi')
hgnc_name_to_id = get_name_id_mapping('hgnc')
hgnc_id_to_entrez_id = get_filtered_xrefs('hgnc', 'ncbigene')
relation_to_adder = {
'controls-state-change-of': BELGraph.add_regulates,
}
namespace_to_dsl = {
'cas': pybel.dsl.Abundance,
'uniprot': pybel.dsl.Protein,
'hprd': pybel.dsl.Protein,
'chebi': pybel.dsl.Abundance,
'hgnc': pybel.dsl.Protein,
}
UNMAPPED = set()
