Multi-objective artificial bee colony for designing multiple genes encoding the same protein

Abstract

The improvement of protein expression levels represents one of the most important goals in synthetic biology. In order to accomplish it, a promising and widely-used strategy lies on integrating multiple genes that encode the subject protein into an organism genome. This important task, however, is affected by several challenging issues. Firstly, the integration of highly similar sequences can potentially induce homologous recombination, a negative effect that implies a reduction in the number of genes effectively integrated. This is the reason why it is important to design multiple protein-coding sequences (also named CDSs) that are as different as possible, between both different CDSs and different subsequences within the same CDS. Additionally, codon usage frequencies in these CDSs should be as highly adapted to the organism as possible. Therefore, this task involves different and conflicting objectives that must be optimized, thus being suitable to be tackled as a multi-objective optimization problem. In this work, we design and implement the algorithm MOABC (Multi-Objective Artificial Bee Colony) to solve the problem of designing multiple CDSs that encode the same protein, considering three objectives to be optimized. The experimental evaluation herein performed suggests that MOABC is able to obtain relevant results, showing statistically significant improvements over the ones found in the literature.