Math Biology Seminar: Simon van Vliet

Microbial communities play major roles in human health and disease and dominate global biochemi­cal cycles. Many proper­ties of these communities emerge from interactions between its mem­ber species and cannot be understood based on the properties of its members living in isolation. The ubiquity of these emergent properties has led some researchers to conclude that microbial commu­nities evolve by multi-level selection (MLS) where selection acts at both the levels of individuals and communi­ties. However, this view is controversial and whether MLS plays a role in natural microbial communities has been heatedly debated in the recent literature. The main controversy in the de­bate is whether community level properties can be inherited. Whether this is the case will depend on the quantitative details of how communities are assembled. For example, the community composition of host-associated communities that are vertically-transmitted between host generations will remain relatively constant through time, allowing for high degrees of heritability of community level traits. In contrast, the composition of communities assembled from the environment can fluctuate strongly in time, leading to low levels of heritability. In general, we need a quantitative method to predict the degree of heritability based on the dynamics of community assembly. Here we present such a method which is based on a recently published MLS framework that explicitly models both individual and community level dynamics. We are attempting to extract generalized rules that can predict the degree of heritability of community level traits for naturally occurring microbial communities. With this framework, we hope to change the focus of the ongoing debate from the question of whether microbial communities can evolve by MLS to the more productive question of which communities can evolve by MLS.