TY - GEN
T1 - A Multi-graph Deep Learning Model for Predicting Drug-Disease Associations
AU - Zhao, Bo Wei
AU - You, Zhu Hong
AU - Hu, Lun
AU - Wong, Leon
AU - Ji, Bo Ya
AU - Zhang, Ping
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - Computational drug repositioning is essential in drug discovery and development. The previous methods basically utilized matrix calculation. Although they had certain effects, they failed to treat drug-disease associations as a graph structure and could not find out more in-depth features of drugs and diseases. In this paper, we propose a model based on multi-graph deep learning to predict unknown drug-disease associations. More specifically, the known relationships between drugs and diseases are learned by two graph deep learning methods. Graph attention network is applied to learn the local structure information of nodes and graph embedding is exploited to learn the global structure information of nodes. Finally, Gradient Boosting Decision Tree is used to combine the two characteristics for training. The experiment results reveal that the AUC is 0.9625 under the ten-fold cross-validation. The proposed model has excellent classification and prediction ability.
AB - Computational drug repositioning is essential in drug discovery and development. The previous methods basically utilized matrix calculation. Although they had certain effects, they failed to treat drug-disease associations as a graph structure and could not find out more in-depth features of drugs and diseases. In this paper, we propose a model based on multi-graph deep learning to predict unknown drug-disease associations. More specifically, the known relationships between drugs and diseases are learned by two graph deep learning methods. Graph attention network is applied to learn the local structure information of nodes and graph embedding is exploited to learn the global structure information of nodes. Finally, Gradient Boosting Decision Tree is used to combine the two characteristics for training. The experiment results reveal that the AUC is 0.9625 under the ten-fold cross-validation. The proposed model has excellent classification and prediction ability.
KW - Computational drug repositioning
KW - Drug-disease associations
KW - Multi-graph deep learning
UR - http://www.scopus.com/inward/record.url?scp=85113778281&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-84532-2_52
DO - 10.1007/978-3-030-84532-2_52
M3 - 会议稿件
AN - SCOPUS:85113778281
SN - 9783030845315
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 580
EP - 590
BT - Intelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings
A2 - Huang, De-Shuang
A2 - Jo, Kang-Hyun
A2 - Li, Jianqiang
A2 - Gribova, Valeriya
A2 - Bevilacqua, Vitoantonio
PB - Springer Science and Business Media Deutschland GmbH
T2 - 17th International Conference on Intelligent Computing, ICIC 2021
Y2 - 12 August 2021 through 15 August 2021
ER -