NEW APPROACHES TO THE MODELING OF SPECIATION

点击翻译按钮获取中文摘要

The overall aim of this project is to better understand biological processes leading to speciation through a study of the dynamic and static behavior of various multilocus models. Existing approaches to the modeling of speciation are deficient in several ways. Usually a small number (one or two) of loci or quantitative traits are considered, selectin is assumed to be weak, population size is considered to be constant, and only a very limited number of selection regimes mostly reflecting the dominant paradigm of "rugged adaptive landscapes" have been studied. Recently a hew metaphor of "holey adaptive landscapes" (Gavrilets and Gravner, 1997) has been put forward as a plausible alternative to the conventional view of rugged adaptive landscapes. This metaphor, which can be traced to a two- locus two-allele model described by Dobzhansky (1937), is the core of the research proposed here. Using a combination of analytical methods and extensive numerical simulations this project will study the dynamics of speciation of holey adaptive landscapes taking into account various factors operating in natural populations. In developing and analyzing speciation models both standard population genetics approaches and novel methods recently developed in physics will be used. This project will investigate the structure of holey adaptive landscapes by focusing on correlated landscapes and concentrating on the possibilities for genetic divergence and speciation introduced by recombination. The dynamics of population evolution on holey adaptive landscapes as determined by mutation, segregation and recombination, random and non-random mating, spatial distribution, population growth and random drift will be studied. Theoretical assumptions and predictions will be tested against properties of natural and laboratory populations. For this purpose, distributions of fitnesses in hybrids and backcrosses, responses of laboratory populations to artificial selection, characteristics of hybrid zones, and macroevolutionary patterns deduced from the fossil record will be used.