Individual-based modelling of adaptation in marine microbial populations using genetically defined physiological parameters

Recent advances in genomics have led to a dramatic upwards revision of marine microbial diversity and a greater appreciation of the important role evolutionary dynamics play in structuring microbial communities. This has presented a significant challenge to marine ecosystem models, which are traditionally diversity poor, and often do not include adaptive/evolutionary processes. Here we explore the use of evolutionary individual-based models (IBMs) as a means of addressing some of these issues. In the IBM, we associate a digital ‘genome’ with each agent, which codes for the phenotypic traits of simulated organisms. Random mutations at the point of reproduction then allow adaptation in response to changing environmental conditions. Trade-offs between different physiological parameters result in different growth strategies emerging under different forcing scenarios. As an idealised test-case we consider resource competition in a chemostat environment, and compare the individual-based approach to a more traditional population-level model. When run in a non-evolutionary context using a clonal population of agents, the IBM reproduces the results of the population-level model. With evolutionary processes enabled, optimally adapted agents are observed to rise to prominence within the agent population. Their physiological trait values are shown to compare well with theoretical optimal trait combinations derived using resource competition theory. In more variable environments, the model is also shown to capture adaptation in response to changed environmental conditions. We conclude that IBMs represent a useful framework for building detailed models linking (sub-)individual-level processes to emergent ecosystem-level behaviour in simplified 0- or 1-D representations of the environment, which should complement global marine ecosystem models of high spatial complexity but necessarily simple process representation.

► We describe an evolutionary IBM of microbial growth.
► The model compares favourably with a more traditional population-level model.
► Emergent properties of model agents agree with theoretical analyses.
► The IBM is able to successfully represent adaptation to changing nutrient regimes.
► The future development of IBMs for marine ecosystem modelling is advocated.