Vagne Constance, David Jacques, Tavaud Muriel, Fontez Bénédicte
Montpellier SupAgro, UMR AGAP, F-34060 Montpellier, France.
Montpellier SupAgro, UMR MISTEA, F-34060 Montpellier, France.
J Theor Biol. 2015 Dec 7;386:44-54. doi: 10.1016/j.jtbi.2015.08.013. Epub 2015 Sep 1.
Since the dawn of agriculture, humans have applied artificial selection on traits of interest, regardless of their genetic architecture. Yet, still today, most models used to study and streamline this process overlook genetic interactions. In this study, we determined the conditions in which a target genotype can be fixed when truncation selection is applied on an epistatic trait. Previous studies have shown that reciprocal sign epistasis with two fitness peaks of unequal height involves multiple equilibrium states, i.e. below one critical parameter value, such as a critical recombination rate, one genotype may be fixed, and above it, another one may be fixed. Using a haploid bi-locus model, we identified which genotype would be fixed, and how quickly, in an infinite population selected for a phenotypic trait subject to reciprocal sign epistasis with unequal peak heights, depending on two criteria: the recombination rate and percentage of selected individuals. The critical parameter values at which bistability sets in, were also calculated. These results were complemented by stochastic simulations in finite populations. Our results confirmed that, in the case of fitness under reciprocal sign epistasis, high recombination rates induce blockage at the local optimum or attainment of an equilibrium state between the two peaks. However, if linkage disequilibrium is negative in the initial population, recombination is necessary to create the most favorable genotype. Therefore, in this case, reciprocal sign epistasis favors non-null recombination rates, particularly if selection is intense.
自农业诞生以来，人类就对感兴趣的性状进行人工选择，而不考虑其遗传结构。然而，直至今日，大多数用于研究和简化这一过程的模型都忽略了基因相互作用。在本研究中，我们确定了对上位性性状进行截断选择时，目标基因型能够被固定的条件。先前的研究表明，具有两个不等高适应度峰值的互反符号上位性涉及多个平衡状态，即低于一个临界参数值，如临界重组率时，一种基因型可能被固定，而高于该值时，另一种基因型可能被固定。我们使用单倍体双基因座模型，根据两个标准：重组率和被选择个体的百分比，确定了在为具有不等高峰值的互反符号上位性的表型性状选择的无限种群中，哪种基因型将被固定以及固定的速度有多快。还计算了双稳态开始时的临界参数值。这些结果通过有限种群中的随机模拟得到了补充。我们的结果证实，在互反符号上位性下的适应度情况下，高重组率会导致在局部最优处受阻或在两个峰值之间达到平衡状态。然而，如果初始种群中的连锁不平衡为负，重组对于创造最有利的基因型是必要的。因此，在这种情况下，互反符号上位性有利于非零重组率，特别是在选择强烈时。
