A Multi-graph Deep Learning Model for Predicting Drug-Disease Associations

Bo Wei Zhao, Zhu Hong You, Lun Hu, Leon Wong, Bo Ya Ji, Ping Zhang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

8 Scopus citations

Abstract

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.

Original languageEnglish
Title of host publicationIntelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings
EditorsDe-Shuang Huang, Kang-Hyun Jo, Jianqiang Li, Valeriya Gribova, Vitoantonio Bevilacqua
PublisherSpringer Science and Business Media Deutschland GmbH
Pages580-590
Number of pages11
ISBN (Print)9783030845315
DOIs
StatePublished - 2021
Externally publishedYes
Event17th International Conference on Intelligent Computing, ICIC 2021 - Shenzhen, China
Duration: 12 Aug 202115 Aug 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12838 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Conference on Intelligent Computing, ICIC 2021
Country/TerritoryChina
CityShenzhen
Period12/08/2115/08/21

Keywords

  • Computational drug repositioning
  • Drug-disease associations
  • Multi-graph deep learning

Fingerprint

Dive into the research topics of 'A Multi-graph Deep Learning Model for Predicting Drug-Disease Associations'. Together they form a unique fingerprint.

Cite this