Chemical Reaction Intermediate State Kinetic Optimization by Particle Swarm Optimization

Large biological molecules such as proteins associating to form multi-component complexes are attracting more and more research interests. The association reaction of the large biological molecules are closely related with associate rate and reaction intermediate states which are key to elucidate the reaction pathways as their kinetic and structural characteristics which shed lights on the reaction process and energy landscape. This paper proposes a novel method modelling the chemical reactions by using neural networks with the help of the predefined chemical reaction model, and then follows by using the typical particle swarm optimization algorithm to minimize the error between the output of neural networks and experimental data. Experiments are conducted to demonstrate the proposed method as a promising way dealing with this difficult task.