Hickey Donal A, Golding G Brian
Department of Biology, Concordia University, 7141 Sherbrooke West, Montreal, Quebec H4B 1R6, Canada.
Department of Biology, McMaster University, 1280 Main St. West, Hamilton, Ontario L8S 4K1, Canada.
J Theor Biol. 2018 Apr 7;442:123-128. doi: 10.1016/j.jtbi.2018.01.018.
Natural selection can act at many loci across the genome. But as the number of polymorphic loci increases linearly, the number of possible genotypic combinations increases exponentially. Consequently, a finite population - even a very large population - contains only a small sample of all possible multi-locus genotypes. In this paper, we revisit the classic Fisher-Muller models of recombination, taking into account the abundant standing variation that is commonly seen in natural populations. We show that the generation of new genotypic combinations through recombination is an important component of adaptive evolution based on multi-locus selection. Specifically, high-fitness genotypes are expected to be absent from the initial population when the frequencies of favorable alleles at the selected loci are low. But as the allele frequencies rise in response to selection the missing genotypes will be generated by recombination. Given recombination, if the average frequency of the favored alleles at the various selected loci is equal to p, then the expected number of favorable alleles per chromosome will be equal to pL, where L is the number of loci. As the value of p approaches unity at the selected loci, the number of favorable alleles per chromosome will approach a value of L, i.e., at the end of the selection process a favorable allele will be found at all loci. In the absence of recombination, however, selection will be limited to the highest-fitness genotypes that are already present in the initial population. We point out that the fitness of such initial genotypes is far less than the theoretical maximum fitness because they contain a favorable allele at only a fraction of the loci. Consequently, recombination acts to unblock the adaptive response to multi-locus selection in finite populations. Using simulations, we show that the sexual population can withstand invasion by newly-arising asexual clones. These results help explain the maintenance of sexual reproduction in natural populations.
自然选择可作用于基因组中的多个位点。但随着多态性位点数量呈线性增加，可能的基因型组合数量呈指数增长。因此，一个有限的种群——即使是非常大的种群——也只包含所有可能的多位点基因型中的一小部分样本。在本文中，我们重新审视了经典的费希尔 - 穆勒重组模型，同时考虑了自然种群中常见的大量现存变异。我们表明，通过重组产生新的基因型组合是基于多位点选择的适应性进化的一个重要组成部分。具体而言，当所选位点上有利等位基因的频率较低时，初始种群中预计不存在高适应性基因型。但随着等位基因频率因选择而上升，缺失的基因型将通过重组产生。给定重组，如果各个所选位点上有利等位基因的平均频率等于p，那么每条染色体上有利等位基因的预期数量将等于pL，其中L是位点的数量。随着所选位点上p的值接近1，每条染色体上有利等位基因的数量将接近L，即，在选择过程结束时，所有位点上都会找到一个有利等位基因。然而，在没有重组的情况下，选择将仅限于初始种群中已经存在的最高适应性基因型。我们指出，这些初始基因型的适应性远低于理论上的最大适应性，因为它们仅在一部分位点上含有有利等位基因。因此，重组有助于解除有限种群中对多位点选择的适应性响应的阻碍。通过模拟，我们表明有性种群能够抵御新出现的无性克隆的入侵。这些结果有助于解释自然种群中有性生殖的维持。
