Neural networks modeling signal responses and taxol production of cultured Taxus chinensis cells induced by bio-elicitor

Abstract

Quantitatively describing the signal transduction process is important for understanding the mechanism of signal regulation in cells, and thus, poses both a challenge and an opportunity for chemical and biochemical engineers. An artificial neural network (ANN), in which we took the signal molecules as neural nodes, was constructed to simulate the generation of active oxygen species (AOS) in Taxus chinensis cells induced by a bio-elicitor. The relative contents of AOS in cells predicted by the ANN model agreed well with the experimental data and three notable stages of AOS increase were observed from the 3D figure of AOS generation. The robustness of AOS trajectories indicated that signal regulation in vivo was an integral feedback control model that ensured the adaptation of Taxus chinensis to environmental stress. The artificial neural network was able to predict taxol production as well as determine the optimal concentration of oligosaccharides needed for it.