Prediction of Inhibition Activity of BET Bromodomain Inhibitors using Grid Search-Based Extreme Learning Machine and Molecular Docking

Background: Inhibition activity of the epigenetic readers, such as bromodomain and extra- C terminal domain protein family, is of high significance in many therapeutic applications due to their ability to regulate gene expression as well as the chromatin structure by binding to acetylysine residues. Objectives: In order to effectively and quickly determine the inhibition activity of these compounds for the desired therapeutic application, this work presents a grid search-based extreme learning machine computational intelligence method through which the inhibition activity of forty different compounds of substituted 4-phenylisoquinolinones was determined. Methods: The prediction and generalization capacity of the developed model were assessed using four different error metrics, which include root mean square error, mean absolute error, mean absolute percentage deviation, and correlation coefficient between the measured values and predicted activities. The lead compound (37), together with a kinase inhibitor, LY294002, and a bromodomain and extra-C terminal inhibitor, CPI-0610, was docked with a bromodomain-containing protein 4 bromodomain 1, 6P05. Results: The developed model performed better than the existing model with percentage improvement of 44.48%, 35.08%, 20.44%, and 1.23% on the basis of mean absolute percentage deviation, mean absolute error, root mean square error, and correlation coefficient, respectively. The lead compound has a better binding score than LY294002 and CPI-0610. Conclusion: Implementation of the developed model would help in searching for anti-inflammatory as well as anticancer agents for effective therapeutic application.