Predictive modeling of drug resistance has become one of the major research topics in recent years. Since the drug development is a complex process, there is a need of using supporting tools, to improve the safety and to reduce the costs of preclinical analysis of the drugs. In order to improve the accuracy in analyzing a big amount of data, where it is not possible to make mathematical calculations for all hidden relations and processes, machine learning and data mining methods are used. Accelerating scientific discovery in chemical compound space is also a motivation for employing intelligent data processing. The main goal of this research is to summarize the methods that are already used for predicting the dynamical interactions on an atomic level that result with drug resistance. Actually, we aim to give the state-of-the-art review of different data driven approaches, which are supported with experimental design. The synthesis of methods presented have shown that data driven approaches are promising tools that can give strong background to the atomic simulations in exponentially growing space of compounds that are synthetically accessible. Furthermore, this paper proves that traditional computational approaches can be upgraded and adapted for effective and efficient predictive modeling of drug resistance.
Predictive Modeling, Drug Resistance, Machine Learning, Data Mining