All evolutionary processes depend on the existence of genetic variation. In our research, we aim to obtain a genome-wide quantitative understanding of how genetic variation is generated and maintained by mutation and natural selection. This in turn allows us to identify components of genetic variation that confer function, and probe how genetic variation contributes to such important evolutionary processes as bacterial adaptation to stress, the emergence of antibiotic resistance, and cancer initiation,

Specific areas of research include:

The evolution of antibiotic resistance in the absence of antibiotic exposure

The emergence of antibiotic resistance is a major threat to human health, and is also one of the best examples of on-going adaptation available for study. A number of studies have demonstrated that exposure to different stresses (e.g. starvation, low pH, high salt concentrations) causes bacteria to more frequently develop antibiotic resistance. Such stress-induced accumulation of resistant mutants occurs even when bacteria have never been exposed to the antibiotics to which they become resistant, and is likely to greatly affect the general dynamics of antibiotic resistance emergence and spread. Currently our understanding of the scope of this phenomenon, of its causes, and of its significance for the spread of resistance within natural bacterial populations is very limited. We are conducting a comprehensive interdisciplinary study, combining microbiology, cutting edge genomic analyses, evolutionary theory, and bioinformatics aimed at systematically addressing these key gaps left in our understanding of the emergence of antibiotic resistance under stress. The results of this project should have far-reaching impact on our understanding of the evolution of antibiotic resistance. In addition, this project should lead to more general insights into the bacterial stress-response, into how natural selection and mutation together drive evolutionary outcomes, and into the manner in which adaptation to a given factor (e.g. antibiotic resistance), may emerge as a result of an unrelated challenge (e.g stress).

Disentangling the effects of mutation and natural selection in introducing biases to patterns of genetic variation

In the past few years we have been witnessing a revolution in genome sequencing technologies. While up until very recently only a handful of sequencing centers had the capability to sequence even single bacterial genomes, it is now possible, even for individual labs, to sequence hundreds of bacterial genomes at relatively low costs. This allows for the generation of whole genome sequence data that can be tailor-made to the needs of individual research groups. The sequencing revolution has also led to a vast increase in the amount of publicly available whole-genome sequence data, most strikingly for bacteria where thousands of genomes are already available, with thousands more in progress. This newly acquired easy access to vast quantities of whole-genome sequence data allows us to identify genetic variation at unprecedented scales. Yet elucidating which components of this genetic variation contribute to function is a far more difficult problem. To address this problem, it is useful to consider how genetic variation is generated and maintained. Genetic variation is generated by mutation. This variation is then either maintained or removed by natural selection and stochastic forces. Because different classes of mutations occur at different frequencies, the mutational process introduces biases into the patterns of genetic variation it generates. Selection introduces further biases into patterns of genetic variation, because different classes of mutations have different fitness effect distributions. Since selection only affects functional variation, understanding which biases result from mutation and which from selection will bring us an important step closer to understanding which parts of genetic variation, and which genetic traits hold functional significance. Biases in patterns of variation generated under severely relaxed selection are introduced mostly by mutation. In our research, we utilize artificially generated, as well as naturally occurring scenarios of severely relaxed selection to characterize mutational biases, and probe how these vary across bacterial species and within populations. This in turn allows us to infer biases to patterns of variation that are selection driven and thus likely functional.

How does selection act on somatic mutations?

Much is known about the extent to which natural selection affects germline mutations in different organisms. At the same time far less in understood about how somatic mutations are affected by selection when they occur in different healthy and malignant tissues. While somatic mutations are not inherited they are nevertheless subject to selection at the cellular level because they can lead to reduction or increases in cellular fitness. At the same time, there is selection at the organismal level to limit somatic evolution, because somatic mutations may be involved in loss of tissue function, in aging, and are involved in both the initiation and progression of cancer. Vast quantities of data have been made recently available, as part of the Cancer Genome Atlas project, of whole-genome, and exome sequences of large variety of tumors and associated healthy tissues. By combining these data with data from RNA-seq studies of gene expression levels in different tissues, we are trying to understand how the effects of purifying and positive selection vary between: (1) different tissue types; (2) healthy and malignant tissues; (3) types of tumors; (4) along stages of cancer progression. Together, the proposed study should lead to important insights into the dynamics of somatic substitution accumulation in different tissues and into how these dynamics change as a function of cancer initiation and progression.

Bacterial genomic variation driven by gene gain and loss events

Bacteria evolve not only through mutations to existing genomic sequences but also through extensive gene gain and gene loss events. As a result most bacterial species are characterized by a pan-genome containing a core of conserved genes and a large set of genes that are present only in a subset of strains. We acquire and combine data of full genome sequences of bacterial strains together with metagenomic data, extracted from the environments in which these strains reside, to study the dynamics of gene loss and gain and the contribution of variation driven by such gain and loss to niche adaptation.