The goal of the project was to investigate the possibility of building a machine learning model that is able to accurately characterize the binding of compounds to drug targets such as androgen receptors.

During this work, we implemented a machine learning model using Python language and modern Python libraries such as TensorFlow and Pytorch. We trained the model with a large number of data related to drug-target interaction taken from Binding DB publicly available databases.

We tested the performance of the machine learning model on the set of selective androgen receptor modulators.

​

A machine learning model for prediction drug–target interaction was developed. The machine learning model was able to predict the efficacy of the SARMs compounds with 74% accuracy. Our data indicate that machine learning models are useful tools for drug development. The technologies that incorporate machine learning and AI have become important tools that can be applied widely in various stages of drug development, such as identification and validation of drug targets, designing of new drugs, drug repurposing, improving the research and development efficiency in the pharmaceutical industry.