Evolutionary Responses to a Constructed Niche: Ancient Mesoamericans as a Model of Gene-Culture Coevolution

Hünemeier, Tábita; Amorim, Carlos Eduardo Guerra; Azevedo, Soledad; Contini, Veronica; Acuña-Alonzo, Víctor; Rothhammer, Francisco; Dugoujon, Jean-Michel; Mazières, Stephane; Barrantes, Ramiro; Villarreal-Molina, María Teresa; Paixão-Côrtes, Vanessa Rodrigues; Salzano, Francisco M.; Canizales-Quinteros, Samuel; Ruiz-Linares, Andres; Bortolini, Maria Cátira

(1) Kimura and Ohta [31] were the first to consider the relations between allele age and its frequency. With this purpose they developed the equation E(t1) = [22p/(12p)]ln(p), where E(t1) = expected age, time is measured in units of 2N generations, and p = population frequency [31] . For ABCA1*230Cys, we considered the average of frequencies of all populations and only those from Mesoamerica/Central America (p = 9.6 and 15.4, respectively; Table 1 ). A generation time of 25 years and N = 720 (number of generation considering the upper limit for the peopling of America, 18,000 YBP [32] ) were assumed. (2) Slatkin and Rannala [33] began to exploit Linkage Disequilibrium (LD) to estimate allele ages, based on variation among different copies of the same allele, where the age of an allele is estimated by the intra-allelic variation following the LD exponential decay due to recombination and mutation rates. Rannala and Reeve [34] , on the other hand, explored the use of LD to map genes, as well as to obtain the allele age using a Markov Chain Monte Carlo framework. We applied this method to obtain a second ABCA1*230Cys age estimative using the DMLE+ v2.2 software (http://www.dmle.org). This program allows a Bayesian inference of the mutation age using an intra-allelic coalescent model to assess LD across the nineteen SNPs that occur around ABCA1*230 (rs2065412, rs2515601, rs2472386, rs2274873, rs2487054, rs4149290, rs2487039, rs2472384, rs2253174, rs2230806, rs2230805, rs2249891, rs4149281, rs4743764, rs1929841, rs2000069, rs2275542, rs3904998, and rs4149268). Taken (c.2) Test to detect deviations from neutrality. Based on the long-range haplotype test [35] and integrated haplotype scores [9, 36] Acuñ a-Alonzo et al. [22] suggested that the autochthonous Native American ABCA1*230Cys allele could have been positively selected. However, demographic events, population structure, and other stochastic processes can create complex patterns in the genome, obscuring signals of natural selection or mimicking adaptive processes [37] . Additionally, positive and balancing selections show different effects on the genetic diversity patterns within and between populations [38] . Therefore, we performed additional analyses to explore these issues and elucidate the factors responsible for the eventual effect of natural selection on the ABCA1*230 locus.