TY - GEN
T1 - Repurposing Drugs for Alzheimer's Diseases through Link Prediction on Biomedical Literature
AU - Xiao, Yongkang
AU - Hou, Yu
AU - Zhou, Huixue
AU - Diallo, Gayo
AU - Fiszman, Marcelo
AU - Wolfson, Julian
AU - Kilicoglu, Halil
AU - Chen, You
AU - Xu, Hua
AU - Mantyh, William G.
AU - Zhang, Rui
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Recently, computational drug repurposing has emerged as a promising method for identifying new interventions for diseases. This study predicts novel drugs for Alzheimer's disease (AD) through link prediction on our developed biomedical knowledge graph. We constructed a comprehensive knowledge graph containing AD concepts and various potential interventions, called ADInt, by integrating a dietary supplement (DS) domain knowledge graph, SuppKG, with semantic triples from SemMedDB database. Four knowledge graph embedding models (TransE, RotatE, DistMult and ComplEX) and two graph convolutional network models (R-GCN and CompGCN) were compared to learn the representation of ADInt. R-GCN outperformed other models by evaluating on the time slice test set and the clinical trial test set, and was used to generate the score tables for the link prediction task. According to the results of link prediction, we proposed candidate drugs for AD. In conclusion, we presented a novel methodology to extend an existing knowledge graph and discover novel drugs for AD. Our method can potentially be applied to other clinical problems, such as discovering drug adverse reactions and drug-drug interactions.
AB - Recently, computational drug repurposing has emerged as a promising method for identifying new interventions for diseases. This study predicts novel drugs for Alzheimer's disease (AD) through link prediction on our developed biomedical knowledge graph. We constructed a comprehensive knowledge graph containing AD concepts and various potential interventions, called ADInt, by integrating a dietary supplement (DS) domain knowledge graph, SuppKG, with semantic triples from SemMedDB database. Four knowledge graph embedding models (TransE, RotatE, DistMult and ComplEX) and two graph convolutional network models (R-GCN and CompGCN) were compared to learn the representation of ADInt. R-GCN outperformed other models by evaluating on the time slice test set and the clinical trial test set, and was used to generate the score tables for the link prediction task. According to the results of link prediction, we proposed candidate drugs for AD. In conclusion, we presented a novel methodology to extend an existing knowledge graph and discover novel drugs for AD. Our method can potentially be applied to other clinical problems, such as discovering drug adverse reactions and drug-drug interactions.
KW - Alzheimer's disease
KW - biomedical knowledge graph
KW - drug repurposing
KW - graph embedding
KW - link prediction
UR - http://www.scopus.com/inward/record.url?scp=85181565046&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85181565046&partnerID=8YFLogxK
U2 - 10.1109/ICHI57859.2023.00137
DO - 10.1109/ICHI57859.2023.00137
M3 - Conference contribution
AN - SCOPUS:85181565046
T3 - Proceedings - 2023 IEEE 11th International Conference on Healthcare Informatics, ICHI 2023
SP - 750
EP - 752
BT - Proceedings - 2023 IEEE 11th International Conference on Healthcare Informatics, ICHI 2023
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 11th IEEE International Conference on Healthcare Informatics, ICHI 2023
Y2 - 26 June 2023 through 29 June 2023
ER -