Prediction of Angiotensin-Converting-Enzyme (ACE) Inhibitors as Antihypertensive Agents Using Artificial Neural Network Optimized by Artificial Bee Colony

This study uses a dataset from the ChEMBL database to predict the activity of Angiotensin-Converting Enzyme (ACE) inhibitors as antihypertensive medicines by use of the Artificial Bee Colony (ABC) algorithm and Artificial Neural Network (ANN). Although they are time-consuming, expensive, and prone to uncertainty, traditional methods include wet-lab testing are often utilized for ACE inhibitor identification. This study intends to increase predictive performance by including systematic optimization through the ABC algorithm with ANN. Architectural and hyperparameters of ANN models were optimized using the ABC algorithm. With an R² of 0.683 on the test set, the model with 10 population size showed the greatest performance among the five models evaluated, therefore proving its efficacy in reflecting the general data pattern. These findings show the possibilities of integrating the ABC algorithm with ANN for precise prediction of ACE inhibitor activity, therefore offering a faster and less expensive substitute for conventional approaches. Larger and more varied datasets are required for more thorough investigation to confirm the generalizability of this method for several drug discovery uses.