Michael Desai
Department of Physics
Harvard University
Genetic variation is essential for a population to adapt to new environments. Yet the adaptation itself destroys this variation. So the amount of variation in evolving populations is set by a balance between creation and distruction of genetic diversity. This balance is essentially between mutation generating variation and selection destroying it -- a classic problem of mutation-selection balance that seems very simple at first.
Despite its apparent simplicity, however, this problem is actually quite subtle and presents a number of interesting theoretical challenges. Most importantly, finite population sizes and fluctuation effects are crucial, in a way which is somewhat unusual in population genetics. In this talk, I'll describe some of these difficulties in more detail and discuss our ways of understanding them. After talking about our solution to the problem in one simple model, I'll talk about further difficulties that arise in other regimes and our preliminary attempts at attacking them.
Why is all this important? The amount of variation that can be maintained (and how this changes with population parameters) is essential in understanding the speed with which populations can evolve, in comparing different experimental and natural populations, and in understanding the advantages provided by various increases in evolvability such as sex or mutator phenotypes.
For those who want more a more detailed outline of the issues and analysis, we have a draft paper (currently in review), available on my website. This describes the theoretical approach in outline and discusses experiments which we have performed to test various of these ideas.