The population genomics of rhesus macaques (Macaca mulatta) based on whole-genome sequences

Abstract

Rhesus macaques (Macaca mulatta) are the most widely used nonhuman primate in biomedical research, have the largest natural geographic distribution of any nonhuman primate, and have been the focus of much evolutionary and behavioral investigation. Consequently, rhesus macaques are one of the most thoroughly studied nonhuman primate species. However, little is known about genome-wide genetic variation in this species. A detailed understanding of extant genomic variation among rhesus macaques has implications for the use of this species as a model for studies of human health and disease, as well as for evolutionary population genomics. Whole-genome sequencing analysis of 133 rhesus macaques revealed more than 43.7 million single-nucleotide variants, including thousands predicted to alter protein sequences, transcript splicing, and transcription factor binding sites. Rhesus macaques exhibit 2.5-fold higher overall nucleotide diversity and slightly elevated putative functional variation compared with humans. This functional variation in macaques provides opportunities for analyses of coding and noncoding variation, and its cellular consequences. Despite modestly higher levels of nonsynonymous variation in the macaques, the estimated distribution of fitness effects and the ratio of nonsynonymous to synonymous variants suggest that purifying selection has had stronger effects in rhesus macaques than in humans. Demographic reconstructions indicate this species has experienced a consistently large but fluctuating population size. Overall, the results presented here provide new insights into the population genomics of nonhuman primates and expand genomic information directly relevant to primate models of human disease.

Figure 1.

(A) Density of autosomal SNVs observed in rhesus (Indian rhesus only, 123 samples) and human (1000 Genomes Project data, phase 1, 123 randomly selected samples). Ind_123 indicates results for 123 rhesus samples; Hum_123 for 123 human samples; Ind_49 indicates 49 low-coverage (average 9.47×) Indian rhesus samples; Hum_49 for 49 human samples with the average coverage 9.44×. (B) Density of autosomal SNVs observed in Chinese rhesus (sample size n = 9), Indian rhesus (nine samples randomly sampled from 123 Indian rhesus samples), chimpanzee (nine samples randomly sampled from 10 samples; downloaded from http://panmap.uchicago.edu), and human (1000 Genomes Project data, phase 1, nine samples randomly sampled from 1092 samples).

Figure 2.

(A) Genome-wide comparison of recombination rates in syntenic regions scaled and averaged over 1-Mb segments in the Indian rhesus genome (blue and cyan, sample n = 49) and human population-average Hapmap genetic map (red and orange). (B) Distribution of 4N_er estimated directly for 534 autosomal syntenic regions that are orthologous in humans and rhesus. (C) Recombination rates scaled by 100 kb in a given syntenic region (human genome coordinate Chr 6: 70,000,000) for humans (orange line) and rhesus (blue line).

Figure 3.

Demographic histories inferred by the stairway plot and PSMC. (A) Stairway plot estimation and 95% CI based on six high-coverage (>20×) Chinese rhesus and 75 high-coverage (>20×) Indian rhesus. (B) PSMC estimations with pattern “6+29*2” for three high-coverage Chinese rhesus (35086, 36013, and 37945) and three high-coverage Indian rhesus (34770, 36461, and 36470): (N_e) effective population size.