Pattern Discovery from High-Order Drug-Drug Interaction Relations

Wen-Hao Chiang; Titus Schleyer; Li Shen; Lang Li; Xia Ning

doi:10.1007/s41666-018-0020-2

Pattern Discovery from High-Order Drug-Drug Interaction Relations

J Healthc Inform Res. 2018 Jun 18;2(3):272-304. doi: 10.1007/s41666-018-0020-2. eCollection 2018 Sep.

Authors

Wen-Hao Chiang¹, Titus Schleyer², Li Shen³, Lang Li⁴, Xia Ning^{1

5}

Affiliations

¹ Department of Computer & Information Science, Indiana University - Purdue University Indianapolis, Indianapolis, IN 46202 USA.
² Center for Biomedical Informatics, Regenstrief Institute, Indianapolis, IN 46202 USA.
³ Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania, Pennsylvania, PA 19104 USA.
⁴ Department of Biomedical Informatics, The Ohio State University, Columbus, OH 43210 USA.
⁵ Center for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indianapolis, IN 46202 USA.

Abstract

Drug-drug interactions (DDIs) and associated adverse drug reactions (ADRs) represent a significant public health problem in the USA. The research presented in this manuscript tackles the problems of representing, quantifying, discovering, and visualizing patterns from high-order DDIs in a purely data-driven fashion within a unified graph-based framework and via unified convolution-based algorithms. We formulate the problem based on the notions of nondirectional DDI relations (DDI-nd's) and directional DDI relations (DDI-d's), and correspondingly developed weighted complete graphs and hyper-graphlets for their representation, respectively. We also develop a convolutional scheme and its stochastic algorithm ${SD}^{2} {ID}^{2} S$ to discover DDI-based drug-drug similarities. Our experimental results demonstrate that such approaches can well capture the patterns of high-order DDIs.

Keywords: Convolution; Drug-drug interactions; Drug-drug similarities; Graph representation; Stochastic algorithm.