Understanding how variations in genotype translate to variation in phenotype is a long-standing problem in molecular biology. We are interested in improving our ability to predict the phenotypic consequences of genetic variants in bacteria. This can be done with a combination of “forward” mechanistic models or more agnostic “reverse” statistical genomics approaches. The lab has access to a large collection of E. coli strains for this exact purpose.

Certain bacterial species such as E. coli have so-called open pangenomes, meaning that new genes - also called “accessory genes” - will be discovered as the genomes of new isolates is being sequenced. Current functional genomics techniques simply cannot scale up fast enough to catch up with the deluge of genes of unknown function that are discovered every day. We are therefore interested in developing methods to assign function to accessory genes in-silico. The wealth of data available for reference species together with advancements in machine learning are a potential way forward to solve this problem.

It has been demonstrated that genetic variability across strains influences gene essentiality profiles. We suspect that these differences might also influence the ability to adapt to a selective pressure, such as the use of antimicrobials. Understanding whether different genetic backgrounds are more or less prone to develop antimicrobial resistance is therefore one of the ways the lab is trying to understand the influence of bacterial pangenomes to clinically relevant phenotypes.