FOR 1078: Natural Selection in Structered Populations

本研究单位的首要目标是研究现实人口环境中的自然选择。由于自然选择作为适应的驱动力，理解自然选择无疑是进化生物学最重要的目标之一。虽然在过去的二十年里，我们对自然选择学到了很多，但最近的研究表明，对选择事件的可靠推断和详细分析，关键取决于我们对发生选择的种群或物种的了解。例如，如果只分析单个（或少数）基因，自然选择、人口细分和人口统计学对基因组的影响是难以区分的。这一困难的产生是因为正向选择和种群规模扩张对多态性模式的影响非常相似，平衡选择和种群细分在有限基因流下的影响也非常相似。然而，随着种群基因组学方法的出现，这种情况得到了极大的改善。这个概念是（强）自然选择通常发生在单个基因上，而群体的结构（细分或人口统计）影响整个基因组。因此，我们从同一生物中调查大量基因座的能力极大地促进了自然种群选择的研究。群体遗传分析的基础模型通常是标准均衡模型（即所谓的Wright-Fisher模型）。然而，大多数自然种群在某种程度上偏离了种群遗传学的标准平衡模型。该模型的一个重要假设是随机交配（panmixia）。然而，正如吉莱斯皮所强调的那样，“如果一个物种的活动范围超过了个体一生中活动的距离，那么它就不可能随机交配。偏离随机交配会对一个物种的进化动力产生深远的影响。”种群可能会以几种方式偏离随机交配。在这个研究股，我们将集中研究人口细分问题。也就是说，我们考虑到本地种群（deme）中的个体是随机交配的，但如果迁移受到限制，deme之间的交配可能会受到限制。Wright-Fisher模型的另一个重要假设是自然种群在很长一段时间内是稳定的。然而，许多调查表明，这一假设经常被违反。因此，在本研究单元中，我们重点关注与赖特-费雪模型的两个重要偏离：人口细分和人口历史。

The overarching goal of this Research Unit is the study of natural selection in realistic population settings. Because of its role as a driving force of adaptation, understanding natural selection is undoubtedly one of the most important objectives of evolutionary biology. Although we have generally learned a lot about natural selection in the past two decades, very recent investigations have shown that reliable inference and detailed analysis of selective events depend critically on our knowledge of the population or species in which selection occurred. For example, the effects of natural selection and of population subdivision and demography on the genome are difficult to distinguish if only single (or a small number of) genes are analysed. This difficulty arises because the effects of positive directional selection and population size expansion on the patterns of polymorphism are very similar, as are those of balancing selection and population subdivision with limited gene flow. However, the situation has drastically improved with the advent of population genomics methods. The notion is that (strong) natural selection usually occurs at individual genes, while the structure of a population (subdivision or demography) affects the entire genome. Thus, our ability to survey a large number of loci from the same organism greatly facilitates the study of selection in natural populations. The model underlying population genetic analyses is usually the standard equilibrium model (the so-called Wright-Fisher model). Yet, most natural populations deviate in some way from the standard equilibrium model of population genetics. One important assumption of this model is random mating (panmixia). As Gillespie emphasises, however, "a species whose range exceeds the distance an individual moves in its lifetime cannot possibly mate at random. Departures from random mating can have profound consequences on the evolutionary dynamics of a species." There are several ways in which a population might deviate from random mating. In this Research Unit, we will concentrate on population subdivision. That is, we take into account that individuals within a local population (deme) mate randomly but that mating between demes may be restricted if migration is limited. Another important assumption of the Wright-Fisher model is that natural populations are stable over long periods of time. However, numerous investigations have shown that this assumption is often violated. Thus, in this Research Unit we focus on two important departures from the Wright-Fisher model: population subdivision and demographic history.