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

摘要: Drug–drug interactions (DDIs) may lead to adverse effects and potentially result in drug withdrawal from the market. Predicting DDIs during development would help reduce costs time by rigorous evaluation of candidates. The primary mechanisms are based on pharmacokinetics (PK) pharmacodynamics (PD). This study examines 2D structural similarities drugs DDI prediction through interaction networks including both PD PK knowledge. Our assumption was that a query (Dq) be examined (De) likely have if network De structurally similar Dq. A describes associations between proteins relating for De. These include target proteins, interacting with enzymes, transporters We constructed logistic regression models using only each Dq results indicated our could effectively predict DDIs. It found integrating similarity scores crucial model performance. In particular, combination target- enzyme-related provided largest increase predictive power. Graphical abstract .