Advances and perspectives in computational prediction of microbial gene essentiality

Abstract

Theminimal subset of genes required for cellular growth, survival and viability of an organismare classified as essential genes. Knowledge of essential genes gives insight into the core structure and functioning of a cell. Thismight lead tomore efficient antimicrobial drug discovery, to elucidation of the correlations between genotype and phenotype, and a better understanding of theminimal requirements for a (synthetic) cell. Traditionally, constructing a catalog of essential genes for a given microbe involved costly and time-consuming laboratory experiments. While experimentalmethods have produced abundant gene essentiality data formodel organisms like Escherichia coli and Bacillus subtilis, the knowledge generated cannot automatically be extrapolated to predict essential genes in all bacteria. In addition, essential genes identified in the laboratory are by definition 'conditionally essential', as they are essential under the specified experimental conditions: these might not resemble conditions in themicroorganisms' natural habitat(s). Also, large-scale experimental assaying for essential genes is not always feasible because of the time investment required to setup these assays. The ability to rapidly and precisely identify essential genes in silico is therefore important and has great potential for applications inmedicine, biotechnology and basic biological research. Here, we review the advancesmade in the use of computationalmethods to predictmicrobial gene essentiality, perspectives for the future of these techniques and the possible practical applications of essential genes.