egKnock: identifying direct gene knockout strategies for microbial strain optimization based on metabolic network with gene-protein-reaction relationships

Abstract

BackgroundGene knockout method has been used to improve the conversion ratio of industrial strains for many chemical products. There are a series of published algorithms to predict the targets for deletion. Based on metabolic networks, many of these algorithms are designed to predict the target of reaction or enzyme deletion. But as for the many-to-many relationship between genes and reactions, reaction or enzyme deletion is not the ideal strategy for metabolic engineering. GDLS algorithm aims to find direct gene deletion target by using local search, but it actually ignores the logic relationship of gene-protein-reaction.\n\nResultsIn this study, we aim to find direct gene deletion targets for metabolic network, but the logic relationship of gene-protein-reaction (GPR) is considered. Our algorithm is call egKnock. At the same time, egKnock will provide the solution with multiple strategies and can maximize the minimum target flux of industrial objective in flux variability analysis. We compare egKnock with the algorithm of GDLS and OptORF by predicting the targets of gene deletion for several chemical products with their flux balance analysis testification, flux variability analysis testification and the main flux distribution.\n\nConclusionsBy comparison with the algorithm of GDLS and OptORF, we can conclude that egKnock is a nice algorithm for identifying direct gene knockout strategies for microbial strain optimization.