Computational Prediction of Protein Subcellular Localization, Genomic Islands, and Virulence to Aid Antigen Discovery

Abstract

Subunits of bacterial proteins involved in virulence, as well as bacterial cell surface and secreted proteins, may be prioritized as potential vaccine components warranting further study. Computational prediction of surface-exposed and extracellular bacterial proteins (that are more accessible to the host immune system) now exceeds the accuracy of most high-throughput “wet lab” methods and has improved markedly in recent years. Identification of proteins encoded within genomic islands is also now much more accurate than a decade ago and can provide insight on gene stability plus aid virulence gene identification. Identifying proteins/genes that are most likely involved in virulence is also desirable, but further improvements in bioinformatics tools for such predictions are needed. This chapter highlights such computational tools currently available to aid the discovery of new vaccine components, how they have improved in recent years, and summarize what is needed in the future to further accelerate vaccine discovery efforts. At minimum, researchers should consider that these recently improved computational methods can now predict more potential vaccine components, and so some bacterial genomes could benefit from reanalysis with these more accurate methods.