Aoki K
Proc Natl Acad Sci U S A. 1986 May;83(9):2929-33. doi: 10.1073/pnas.83.9.2929.
A stochastic model of gene-culture coevolution, suggested by the "culture historical hypothesis" of Simoons and McCracken, is presented. According to this hypothesis, adult lactose absorption, believed to be an autosomal dominant trait, attained a high frequency in some human populations due to the positive selection pressure induced by culturally determined milk use in those populations. Two-dimensional Kolmogorov backward equations with appropriate boundary conditions are derived for the ultimate fixation probability of milk users, of the gene for adult lactose absorption, and of both jointly, and for the average time until fixation of the gene. These boundary value problems are solved numerically by the Gauss-Seidel method. I define a theoretical measure of the correlation between gene and culture in terms of the three ultimate fixation probabilities. Monte Carlo simulations are conducted to check and extend the numerical results and also to obtain the first arrival time at gene frequency 0.70, which is approximately the highest observed frequency in any population. Two results that pertain to the culture historical hypothesis are obtained. First, the incomplete correlation observed between adult lactose absorption and milk use does not necessarily constitute evidence against the hypothesis. Second, for the postulated genetic change to have occurred within the 6000-year period since the advent of dairying, either the effective population size was of the order of 100, or, if it was of larger order, the selection coefficient probably had to exceed 5%.
本文提出了一个由西蒙斯和麦克拉肯的“文化历史假说”所启发的基因 - 文化共同进化的随机模型。根据这一假说,成人乳糖吸收被认为是一种常染色体显性性状,由于在某些人群中文化决定的奶类食用所引发的正选择压力,该性状在这些人群中达到了高频率。针对奶类使用者、成人乳糖吸收基因以及两者共同的最终固定概率,以及基因固定前的平均时间,推导了带有适当边界条件的二维柯尔莫哥洛夫向后方程。这些边值问题通过高斯 - 赛德尔方法进行数值求解。我根据这三个最终固定概率定义了一种基因与文化之间相关性的理论度量。进行了蒙特卡罗模拟以检验和扩展数值结果,同时还获得了基因频率达到0.70的首次到达时间,这大约是在任何人群中观察到的最高频率。得到了两个与文化历史假说相关的结果。第一,在成人乳糖吸收和奶类食用之间观察到的不完全相关性不一定构成反对该假说的证据。第二,对于自乳业出现以来的6000年时间内发生的假定基因变化而言,要么有效种群大小约为100,要么如果种群大小更大,选择系数可能必须超过5%。