Predicting drug–drug interactions through drug structural similarities and interaction networks incorporating pharmacokinetics and pharmacodynamics knowledge

Authors

• T Takeda
• M Hao
• T Cheng
• SH Bryant
• Y Wang

Abstract

Drug-drug interactions (DDIs) can cause serious adverse effects and sometimes lead to drug withdrawal from the market. During drug development, the prediction of such DDI would help reduce the time and costs by prioritizing drug. The primary mechanisms of DDIs are based on pharmacokinetics (PK) and pharmacodynamics (PD). This study examines the effects of structural similarities of drugs on DDI. We proposed models for predicting DDIs using the structural similarities of drugs from the PK and PD networks and investigated the factors influencing DDIs for further improvement of the predictions. Our assumption was that a query drug (Dq) and a drug to be examined (De) likely have DDI if the drugs in the interaction network of De are structurally similar to Dq. A network of De describes the associations between the drugs and the proteins relating to PK and PD for De. These include target proteins, proteins interacting with target proteins, enzymes, and transporters for De. We constructed logistic regression models for DDI prediction using only 2D structural similarities between each Dq and the drugs in the network of De. The results indicated that our models could effectively predict DDIs. Our work demonstrated: (1) structural similarities between Dq and the drugs in the network of De can be used for predicting DDIs between Dq and De; (2) integrating both structural similarity scores relating to PK and PD was crucial for DDI prediction; (3) including pharmacogenetically associated knowledge only made minor contribution to DDI predictions.

Scientific Focus Area: Computational Biology

This page was last updated on Friday, March 26, 2021