Using hybrid GA-ANN to predict biological activity of HIV protease inhibitors

The prediction of biological activity of a chemical compound from its structural features, representing its physico-chemical properties, plays an important role in drug discovery, design and development. Since the biological data is highly non-linear, the machine-learning techniques have been widely used for modeling it. In the present work, the clustering, genetic algorithm (GA) and artificial neural networks (ANN) are used to develop computational prediction models on a dataset of HIV protease inhibitors. The hybrid GA- ANN technique is used for feature selection. The ANN-QSAR prediction models are then developed to link the structures to their reported biological activity. These models can be useful for predicting the biological activity of new untested HIV protease inhibitors and virtual screening for identifying new lead compounds.