Linear and nonlinear modeling of protein kinase B/AkT

Abstract

AkT is the main protein which was frequently occurred in the human cancer. There are several pathways of AkT which were shown in this paper which lead to cell death/survival. In this paper, we have used the mathematical analysis (linear modeling and non linear modeling) to make a best model of the survival/death proteins, i.e., epidermal growth factor, tumor necrosis factor, and insulin using ten different combinations. The model was made using linear modeling using different regression analysis techniques in which different parameters like mean sq error, root mean sq error, mean abs error, relative sq error, root relative sq error, and relative abs error were analyzed. Later on, Kolmogorov–Smirnov, Anderson–Darling, and chi-square tests were done using different distribution functions. Results with half-normal distribution function are the best as their AD and chi-square values are the maximum. Nonlinear modeling was done using neural network with MLP and RBF approaches.