Toxicity Predicition of 3D-Domain Swapped Predicted Proteins using Machine Learning Approach

In this study efforts are put to createimplimention of development of a structure activity relationship (SAR) based model, that is used to anticipate toxicities and save time, testing on animals, money, chemicals at the time of protein-ligand interaction for drug development. Machine learning models are implemented on seven 3D-domain swapped proteins. In first level classification model were implemented and in second level regression model were implemented. Training dataset includes swapped proteins structure with their Pharmocological activities and contain about 501 features. Because of the extreme imbalance and abundance of characteristics in our dataset, we first carried out feature selection.then the data-preprocessing. We also implemented the K-folds cross validation for the same and the simulation results were obtained.