UNDERSTANDING BEHAVIOR OF ANTINEOPLASTIC MOLECULES BASED ON MLR MODELS

New statistic based models provide a wide area of prediction equipments for different science areas. Among these fields biology have just entered the contest of interdisciplinary sciences. Drug discovery is a long and expensive process which could be decreased with theoretical approaches. In this study, 500 reported assayed anti cancer molecules were extracted from Science Direct articles, sketched, optimized with AM1 basis set with Gaussian, and analyzed for their bioactivities with PASS. Molecular and physicochemical descriptors such as MW, H Donor/Acceptor, LogP, Parachor, and etc of 15 screened compounds were calculated with Dragon and applied for Stepwise Multi Linear Regression process within SPSS. Calculations present that H Donor, H Acceptor, Polarizability, and Refractivity descriptors are the most impacting factors on emergence of antineoplastic activity with higher than 91% accuracy. A valid model is achieved which will be utilized for future molecule discovery and prediction.