Evolutionary algorithm for metabolic pathways synthesis

GERARD, M.; STEGMAYER, G.; MILONE, D.H.

Simposio; XVII Simposio Argentino de Inteligencia Artificial - 45 Jornadas Argentinas de Informática; 2016

Several methods to automatically search for metabolic pathways among compounds have been developed in last years. They are mainly based on classical search algorithms, such as breadth-first and depth-first search, and the A* algorithm. They first model data as an appropiate graph, and then perform the search of a pathway between compounds. Although they have served as a first attempt to address the problem, these methods have at least one of the following limitations: i) they do not consider all the substrates for the reactions (feasibility of reactions is not guaranteed); ii) they cannot model and build branched metabolic pathways; iii) they only search pathways between two compounds;iv) previously synthesized compounds in the reactions chain are not taken into account to select a new reaction (just the last one). Frequently, it leads to find solutions without biological sense. We propose a new algorithm based on evolutionary computation for searching branched metabolic pathways, that links simultaneously several compounds through feasible reactions. This proposal, named evolutionary metabolic synthesizer (EvoMS), uses the concept of expanded set of compounds (ESC) to model metabolic pathways.