Histone dacetylases (HDACs) are a group of enzymes that remove acetyl groups from histones and regulate expression of tumor suppressor genes making them a promising therapeutic target for treatment of cancer by developing a wide variety of inhibitors. Developing these inhibitors requires accurate understanding of how their molecular structures are link to their respective inhibitory properties. A Genetic Function Approximation based Multi-linear regression Quantitative structure activity relationship modelling was performed on a data set of 29 HDAC inhibitors using Semi-empirical (PM3) computational level of theory. The best QSAR model reveals that FMF, Kier3, n5HeteroRing, globaltopo and Kier1 descriptors have pronounced influence on the HDAC inhibitory properties of the compounds. The validation parameters of the best model are LOF = 0.137, R2 = 0.933, R2adj = 0.902, Q2LOO = 0.841, F-value = 30.239, R2pred. = 0.6495. The wealth of information provided by this model will undoubtedly be of immense help in the structural modifications of the studied molecules as a guide to discover additional HDAC inhibitors with greater therapeutic utility.
