Fuzzy multiobjective hierarchical optimization with application to identify antienzymes of colon cancer cells

Abstract

BackgroundDeveloping a new drug from conception to launch is costly and time consuming. Computer-aided methods can reduce research cost and accelerate the development process during the early drug discovery and development stages. MethodsThe present study proposed a fuzzy hierarchical optimization framework for identifying potential anticancer targets in genome-scale metabolic models. In this framework used for evaluating treatment side effects, the mortality of treated cancer cells, viability and metabolic deviation of perturbed normal cells were assigned fuzzy objective functions. Fuzzy set theory and the nested hybrid differential evolution (NHDE) algorithm were applied to solve a maximizing decision-making problem transformed from the hierarchical optimization problem. We verified the existence and limitation of this transformation. Significant FindingsThe NHDE algorithm in the optimization framework identified 12 one-target genes with high hierarchical fitness scores by using a small population size. The experimental data accessed from the DepMap database indicated that most of the identified targets can cause most cell lines to die (except EBP, LSS, and NSDHL). Two-target combinations resulted in higher cell viability and lower side effects than did one-target treatments. However, three- and four-target combinations did not significantly improve cell viability and reduce side effects relative to the two-target combinations.