Bayesian models for long-range population genetics

点击翻译按钮获取中文摘要

Two quite different perspectives can be taken on the problem of species evolution. The historical perspective consists in reconstructing the evolutionary history of all living species (the tree of life), based on fossil evidence and molecular sequence comparisons. This is the task of phylogenetics. Recently, it has motivated important statistical efforts so as to provide reliable tree reconstructions. The demographic and mechanistic perspective, on the other hand, is more concerned with understanding how the intricate interplay between mutation, selection and random drift determines the mode of evolution of extant species, down to the molecular level. It also tries to understand how the ecological factors experienced by a species (such as its population size) influence this interplay. Population genetics provides the theoretical framework for such an understanding. Classical population genetics achieves its aims mainly by observing the distribution of intra-specific genetic diversity (polymorphisms). However, it is progressively understood that very valuable information can also be derived from the large-scale comparative analyses performed by phylogeneticists. This information is less detailed than that derived from polymorphisms. However, it reaches much further back in time, raising the exciting possibility that by reading it out, one could reconstruct the evolution of essential ecological factors experienced by species along the tree of life.