论文部分内容阅读
OBJECTIVE To construct an integrative database for multi-compound drug discovery.METHODS We designed and constructed a database system,which integrates traditional herbal medicine,functional food,and drug combination information.Our database consists of six entity tables,namely drug combinations,functional foods,prescriptions,herbs,compounds and phenotypes.We established strategies for data integration and entity resolution to facilitate heterogeneous information of multi-compound therapies.To standardize the data,instances of entity tables are mapped to international identifiers,and phenotype terms in narrative text are extracted by using the named entity recognition(NER)method.RESULTS The database integrates therapeutic information of traditional herbal medicine,functional foods and combination drugs which is acquired from Traditional Chinese Medicine Information Database(TCM-ID),Food and Drug Administration(FDA)and Drug Combination Database(DCDB).The herb information is mapped to NCBI taxonomy identifiers,and compound information is mapped to PubChem and ChEMBL identifiers for standardization.We also applied MetaMap,a tool for recognizing UMLS concepts from narrative text,to extract phenotype terms.The current version of the database contains 6 291 drug combinations,1 615 functional foods,20 091 prescriptions,8889herbs,227 636 compounds and 11 744 phenotypes.CONCLUSION Our database provides various therapeutic information of multi-compound therapies which serve as a fundamental resource for the polypharmacology research.
OBJECTIVE To construct an integrative database for multi-compound drug discovery. METHODS We designed and constructed a database system, which integrates traditional herbal medicine, functional food, and drug combination information. Our database consists of six entity tables, ie drug combinations, functional foods , prescriptions, herbs, compounds and phenotypes. We established strategies for data integration and entity resolution to facilitate heterogeneous information of multi-compound therapies. To standardize the data, instances of entity tables are mapped to international identifiers, and phenotype terms in narrative text are extracted by using the named entity recognition (NER) method. RESULTS The database integrates therapeutic information of traditional herbal medicine, functional foods and combination drugs which is acquired from Traditional Chinese Medicine Information Database (TCM-ID), Food and Drug Administration (FDA) and Drug Combination Database (DCDB). The herb information is mapped to NC BI taxonomy identifiers, and compound information is mapped to PubChem and ChEMBL identifiers for standardization. We also applied MetaMap, a tool for recognizing UMLS concepts from narrative text, to extract the phenotype terms. The current version of the database contains 6 291 drug combinations, 1 615 functional foods, 20 091 prescriptions, 8889 herbs, 227 636 compounds and 11 744 phenotypes.CONCLUSION Our database provides various therapeutic information of multi-compound therapies which serve as a fundamental resource for the polypharmacology research.