Protein-Drug Binding Analysis Using Change Detection on Binding Site Images

With the rapid increase in the amount of accessible chemical and biological data over the last decade, the use of intelligent methods in drug design and development fields has gained importance. The purpose of this study is to develop a modeling method to use when predicting the binding affinity in protein-drug interactions. In the first step of this two-phased method, the aim was to detect changes in the protein-drug interaction images. Images of protein's binding area, which were obtained when the area was empty and when the area was occupied by drug molecules, were compared. Then, the most changed regions were detected and converted into feature vectors. In the second step, by using the Support Vector Machine Regression algorithm, the binding affinities of each protein-drug complex were predicted and the performance of the method was compared with the observed binding affinity values.