Computational analysis of gene clusters from antimicrobial producing and pathogenic bacteria. (360G-Wellcome-109164_Z_15_A)

This project will develop a novel method for the in silico prediction of the function of bacterial genes based on the analysis of their genomic neighbourhood, with a specific focus on genes of interest in infection biology. Indeed, bacterial genes involved in the same cellular "macro-function", such as a signalling or biosynthetic pathway, are often physically clustering together on the genome. The computational analysis of these gene clusters could help to uncover the role of currently uncharacterised genes in certain pathways. The prediction method will be developed and tested on a dataset of publicly available bacterial genomes. Then, it will be applied to the analysis of clusters putatively involved in the synthesis of peptide antimicrobials or in processes targeted by known antibiotics. Orthologous clusters will be compared to identify protein domains that tend to co-occur in similar clusters ("domain correlation"). However, sometimes two alternative domains may fulfil the same biochemical function. In such a case, these two domains may be present as alternatives within functionally related clusters. This results in an "anti-correlation" of these domains within that cluster. The final aim of this analysis will be the identification of new antimicrobial processing enzymes and possible novel drug targets.