Population Genetic Statistics from Rhesus Macaques (Macaca mulatta) in Three Different Housing Configurations at the California National Primate Research Center

Abstract

This study analyzed the genetic composition of 3382 genetically characterized and pedigreed animals currently maintained under 3 different housing configurations at the California National Primate Research Center, including the indoor colony, outdoor ‘corn cribs,’ and half-acre field cages. Summary statistics based on 15 short tandem repeats strongly suggest significant effects of genetic drift, including the loss of allele diversity, among the enclosures within the housing facilities even though gene flow among the different housing units is actively promoted by colony management. Management methods of selectively harvesting female macaques to prevent overrepresentation of one or only a few matrilines and cross-fostering 1-wk-old infants among breeding cages and corn cribs have been insufficient to prevent genetic subdivisions among the cages and corn cribs and to evenly distribute genetic diversity throughout the colony. In addition to promoting several colony management strategies recommended herein to effectively curb inbreeding and genetic differentiation, current attempts of infant cross-fostering and minimizing matriline fragmentation should be expanded. The inclusion of inbred or highly genetically homogeneous animals with diminished allele diversity in linkage and association studies will likely compromise the potential for identifying allele–disease associations, whereas the inclusion of macaques from different geographic origins or their hybrids (or both) in experimental research confounds interpretations of phenotypic differences, due to inflation of the genetic contribution to phenotypic variance.

In 1962, the California National Primate Research Center (CNPRC) was established to conduct biomedical research dedicated to improving human and animal health and to serve as a resource for nonhuman primate research stock. Research performed at the center provides crucial knowledge on infectious diseases such as AIDS,¹ development of new vaccines,³⁴ xenotransplantation,² reproductive sciences,¹² neurobiology, including cognitive function and neurodegenerative conditions such as Alzheimer disease,²⁷ respiratory and pulmonary disorders, such as asthma,^35,40 biobehavioral organization, behavioral development and social relationships,^4-7 and primate physiology, including therapies for chronic colitis and nutritional deficiencies.

Tracking and registering an animal's location history, use, parents and offspring, acquisition and origin, removal, demography and ancestry, health and reproductive status, and so forth are all fundamental requirements, particularly in large facilities such as the CNPRC because of the size of its captive population, different styles of housing enclosures, and the ever-increasing complexity of demographic and pedigree histories as the animals reproduce. Because the rhesus macaques at the CNPRC comprise a closed colony of animals that resembles a small isolated population, complex patterns of genealogic relationship among individual macaques must be considered to optimally manage genetic variability and population structure of the colony. The advantage of a large, genetically characterized, and well-managed colony, like that at the CNPRC, is that it can supply the research community with sufficient numbers of healthy animals that have known life histories, pedigree affiliation, and demographic and genetic profiles.

Researchers using CNPRC animal resources may find it helpful to understand the genetic impact of colony management decisions on the genetic structure of captive colonies and their implications for biomedical research. For instance, genetic drift causes genetic differences among the different housing enclosures of a captive colony that is closed to gene flow, thereby reducing genetic diversity and hampering investigations of diagnostic allele–marker associations. At the other extreme, increased genetic heterogeneity, phenotypic variance due to genotypic variance, outbreeding depression, and linkage disequilibrium are created when animals whose ancestors originated in different geographic regions (for example, Indian and Chinese rhesus macaques) are bred.^19,22,30

Of the 5200 monkeys presently maintained in quarantine, breeding, and holding facilities of the CNPRC, more than 5000 are rhesus macaques (Macaca mulatta). The indoor rhesus population comprises 2000 animals that have been assigned to specific research projects, ranging from time-mated breeding for reproductive studies to those requiring Level 2 and Level 3 biosafety containment for studies of infectious diseases. The indoor colony also includes nursery-reared, weanling, and aged (older than 18 y) animals. Although the proportion of animals assigned to terminal studies and short-term projects annually has remained relatively constant, the composition of animals in some of the indoor enclosures, such as the hospitals and exposure facilities, is transient and subject to daily, weekly, and monthly changes.

A total of 21 half-acre field cages, each of which accommodates between 80 and 120 macaques, house the CNPRC's long-term breeding resource of approximately 2400 animals. In addition to being a source for infants, juveniles, adults, and geriatric animals, the field cages provide a focus for social behavior studies. The tenure of these field-cage populations as a breeding resource varies in length from 1 y (recent field-cage formations) to 20 y. The live-birth rate in the field cages reaches 78% annually, with an average production of 551 newborns per year. As the number of field cage occupants exceeds its optimal density, surplus animals are used to populate new production field cages. In addition to annual harvest for assignment to research projects, animals are relocated for social reasons (including deleterious aggression, matriline fragmentation, balancing sex ratio, and group stability) and to maximize gene flow among field cages and ‘corn cribs.’²⁰

Animals targeted for removal due to density issues in field cages are high-ranking juvenile males because they, with support from their families, instigate deleterious aggression. An excessively low sex ratio can trigger deleterious aggression among female macaques because females compete over males, and there are too few males to police the group adequately. In addition, matriline fragmentation or patterns of relatedness among individuals in a matriline can affect patterns of aggression and social group stability. As such, some animals of the fragmented matriline are removed to stabilize the cage. Animals are not moved between breeding groups if it is suspected that they would be wounded or killed by other animals. Similarly, higher-ranking animals and those that play key roles in stabilizing the group are not relocated because their removal can result in social overthrow that sometimes can require cage disbanding.

Approximately 800 rhesus macaques are housed in 40 conical-shaped pens called corn-cribs, each of which accommodates 12 to 20 animals. These enclosures provide options for smaller social groups, including young, weaned, animals in peer-housing situations with an adult animal and harem breeding groups that comprise a single male and multiple females in estrous who typically will experience their first birth. Expanding colonies of SPF rhesus macaques (animals that are antibody-negative to cercopithecine herpesvirus 1 [B virus], SIV, type D simian retrovirus, and simian T-lymphotropic virus; 743 macaques) and pure-Indian (302 monkeys) and pure-Chinese (95 macaques) ‘superSPF’ rhesus macaques (animals that are also free of rhesus cytomegalovirus, rhesus rhadinovirus, and simian foamy virus) at the CNPRC are supported by the National Center for Research Resources (NCRR).²⁰ The SPF and superSPF animals range in age from newborn to adult; are housed in designated field cages, corn cribs, or indoor housing; and require separate hospital areas to prevent cross-contamination and to preserve their SPF status.

This study compares the genetic structure and population characteristics of 3382 rhesus macaques in the CNPRC colony living in single-cage indoor housing, half-acre outdoor field cages, and corn cribs. These animals represent multigenerational pedigrees that include great-great-grandparents, great-grandparents, grandparents, parents, and offspring.The complex demography of the study populations over a period of 4-7 generations complicates inferences about large-scale patterns of multigenerational genetic stratification based solely on pedigree information. It has been shown that population-level genome-wide comparisons based on autosomal markers such as microsatellites or short tandem repeats (STR) provide reliable estimates of population stratification and dynamics.^14,25 Therefore, the current study directly assesses the population genetic structure and dynamics based on the STR genotypes that were generated for parentage testing and genetic management at the CNPRC and combines inferences based on population genetics with those based on demographic and pedigree records.

Materials and Methods

Samples.

The rhesus macaque colony at the CNPRC is managed in accordance with IACUC regulations that prescribe the humane care and use of laboratory animals. Almost all living rhesus macaques at the CNPRC have been genotyped and parentage-tested. Among the 3382 macaques studied, 1032 are housed in 10 different indoor colony housing areas, 353 in 40 outdoor corn cribs, and 1997 in 20 outdoor field cages (referred to as NCs 1 to 20); these macaques comprise 52%, 44%, and 83%, respectively, of the total number of animals currently living in each of these 3 housing facilities. The indoor housing areas include: modular buildings, the animal wing, the 4 subdivided quadrants of the Butler Building, the exposure facility, hospitals, a facility containing testing rooms for neurobehavioral studies, the North Wing, the South Wing, the Speed Space, quarantine rooms, and a temporary housing pad for animals transferred from outdoors while they are screened for Shigella before being assigned cages in specific indoor rooms. This space also is used to house outdoor animals assigned to specific projects for 1 or 2 d while undergoing project-related procedures before being returned to their respective outdoor cages. Similarly, quarantine rooms can be used for regular housing for various projects when no pending import shipments that require their use are foreseen.

Like the outdoor field cages, most of the corn cribs house multiple adult male and female macaques (that is, group, gang, or troop caging). Corn cribs CC602, CC807, and CC1004 contain harems, whereas CC202 and CC403 house only adult female rhesus macaques. Table 1 presents the ancestry of the animals housed in each field cage. Outdoor field cage NC17 houses only pure-Chinese animals, whereas NC2, NC5, and NC15 house only pure-Indian animals. All other field cages accommodate varying numbers of Chinese–Indian hybrid animals comprising between 3% (NC10) and 71% (NC18) of the field cage's population. Field cages NC1, NC4, NC19, and NC20 house SPF animals, whereas superSPF animals are housed in NC2, an all-Indian field cage, and NC3, 10% of whose residents are hybrid animals; the pure-Chinese individuals housed in NC17 are all superSPF animals.

Table 1.

Estimates of F-statistics for the different housing areas in the indoor colony, corn cribs (CC), and field cages (NCs)

Colony-wide genetic and pedigree data from December 2008 were used in this study. DNA extraction from whole-blood samples, STR fragment analysis and parentage assignment based on STRs (see reference 24 for a complete description of these markers) were performed by the Veterinary Genetics Laboratory (University of California, Davis, CA). Confirmation of maternity and paternity were based on the purported parents’ genetic profiles, one or the other purported parent's profiles, or the individual animal's genotype record only. After alleles shared between the dam and offspring were ascertained, purported sires that did not possess the implied paternal alleles across all loci were excluded as the sire of the offspring. When the genetic information of the alleged dam was unavailable and the sire unknown, parentage was assumed based on behavioral and circumstantial evidence, including the presence of the purported parents in the same cage during the time of conception of an individual animal and the identity of the female adult with whom an individual animal secured maternal attention and affection as an infant. Parentage that was verified by genetic screening was indicated in the animal records.

Statistical analysis.

All 3382 animals included in the present study have at least 95% complete genotype information across the 15 STR loci studied. These animals represent approximately 80% of the 4300 genetically characterized and pedigreed rhesus macaques currently maintained at the CNPRC.

The exact probability test in GENEPOP²⁹ was used to test for the presence of linkage disequilibrium (the nonrandom association of genotypes occurring at different loci) between pairs of the 15 STR loci. To test the null hypothesis that genotypes at one locus segregate independently of genotypes at any other locus at the 0.05 and 0.01 levels of probability, unbiased estimates were made through randomization (1,000 iterations), and the Markov chain method was used to create a contingency table representing the random association of genotypes at all possible pairs of loci. In addition to assessment of linkage disequilibrium between loci, Hardy–Weinberg equilibrium within each locus was analyzed by using GENEPOP.²⁹

An assignment analysis using the program STRUCTURE^14,28 based on the allele frequencies at 15 STR loci was conducted at sweeps of 500 iterations after a burn-in period of 50 × 10³ to estimate the degree of Chinese or Indian ancestry of the macaques in each of the indoor and outdoor enclosures. Analyses were conducted without a priori ancestry information but with the assumption of K values of 2 regional populations, that is, for Chinese or Indian ancestry. Depending on its STR allele frequencies, an animal was characterized as Chinese or Indian with probability Q, the proportion of its genome that is Chinese or Indian in origin. Because hybrid rhesus macaques will exhibit Chinese and Indian nuclear admixture, that is, both Chinese and Indian STR alleles, the STRUCTURE program facilitates the genetic identification of hybrid individuals¹⁹ and provides a rough estimate of their proportions of Indian and Chinese ancestry.

The F statistics (F_is, F_st, and F_it)³⁸ feature of GENEPOP²⁹ was used to assess the breeding structures in the field cages and corn cribs and their effects on the genetic structure of the different housing facilities studied. F_is estimates the contribution of nonrandom mating, including inbreeding and population substructure, within the different enclosures, whereas F_st estimates genetic differentiation (genetic subdivision) among enclosures, and F_it takes into account the effects of both nonrandom mating and genetic subdivision. Positive F-statistic values indicate an excess of homozygous genotypes relative to gamete frequencies, due to nonrandom mating and genetic subdivision. F-statistics values ranging from 0 to 0.15 and from 0.15 to more than 0.25 imply negligible to moderate and high to very high, respectively, levels of homozygosity. GENEPOP²⁹ was used to estimate the level of pairwise genetic differentiation among and within the different types of enclosures based on Weir and Cockerham pairwise F_st.³⁶ Parameters of genetic diversity, including actual and effective allele numbers, observed heterozygosity, and gene diversity (heterozygosity expected under equilibrium conditions), of each enclosure were computed by using PopGene (version 1.32).³⁹

Results

Genetic structure.

The highest pairwise linkage disequilibrium was observed among the animals in field cages, for which 67 and 94 pairs of loci of the possible 105 pairwise comparisons were significantly associated at the 1% and 5% probability levels, respectively. Among the corn crib population, 29 and 48 pairs of loci were in linkage disequilibrium at P ≤ 0.01 and P ≤ 0.05, respectively whereas 40 and 66 pairs of loci were significantly associated at the 1% and 5% levels of probability among macaques in indoor housing (data not shown). Three pairs of physically linked STRs that are located on the rhesus chromosome 6—D6S276, D6S291, and D6S1691—did not exhibit greater statistically significant association of genotypes at P ≤ 0.05 or P ≤ 0.01 than did unlinked loci. Therefore, despite their synteny, the linked STRs on chromosome 6 were treated similarly to the unlinked markers and were included in all subsequent population genetic analyses. Highly significant (P ≤ 0.01) within-loci Hardy–Weinberg disequilibria in all 3 housing facilities (data not shown) were present, suggesting nonrandom associations among alleles at loci analyzed here.

Based on STRUCTURE analysis, some of the enclosures (for example, NC17, which houses pure-Chinese animals) contained only animals of Chinese origin. However, a large fraction of the animals in all 3 housing facilities exhibited varying proportions of Chinese–Indian ancestry according to STRUCTURE analysis, in close agreement with demographic records of animals with known ancestry (Figure 1).

Figure 1.

A rhesus macaque, represented by a single vertical bar, is characterized as Chinese or Indian with probability Q, the proportion of its genome that is in Chinese (red) or Indian (green) in origin. A hybrid animal will exhibit both the red and green colors depending on its degree of Chinese or Indian ancestry. This analysis was based on 15 STRs. Indoor (AB, modular buildings; AW, animal wing; BB, Baker Building; EX, exposure facility; HO, hospitals; MB, neurobehavioral testing rooms; NW, North Wing; QU, quarantine rooms; SS, Speed Space; SW, South Wing; THPAD, temporary housing pad), n = 1032; corn crib, n = 353; field cages (NC), n = 1997.

Summary F-statistics (Table 1) revealed moderate (indoor housing and field cages) to high (corn cribs) levels of drift-driven differentiation among the housing units of all 3 housing facilities at the CNPRC. The F_st values among the different indoor housing areas and outdoor field cages were 0.04 and 0.06, respectively, whereas differentiation among the corn cribs (F_st = 0.18) was much greater. However, pairwise F_st values among the 3 different housing facilities were negligible (indoor housing versus corn cribs, 0.001; indoor colony versus field cages, 0.0006; corn cribs versus field cages, 0.0017), indicating that genes are distributed randomly among, but not within (as suggested by the range of 29 to 67 [P ≤ 0.01] and 48 to 94 [P ≤ 0.05] pairs of loci that were not in equilibrium), each of the 3 facilities.

Computation of F-statistics (Table 1) showed that 7 of the 11 housing areas within the indoor colony (modular buildings, animal wing, exposure facility, hospitals, North Wing, Speed Space, and temporary housing pad) exhibited very low to moderate levels of nonrandom mating (mean, 0.03), with F_is values ranging from 0.01 to 0.04, whereas the remaining 4 areas exhibited F_is values approaching 0.00. The Speed Space, which includes 37% of the entire indoor colony sampled, exhibited the highest F_is value (0.04). The indoor colony comprises an artificial population of animals selected for projects and does not reflect any long-term breeding strategy. Hospital-housed macaques have a ‘home cage’ to which they typically return within 1 d to 3 mo if they survive. Many of these animals are on terminal experiments and will have no genetic impact on the colony. F_is estimates revealed no evidence of nonrandom mating in the corn cribs (mean, –0.13) and field cages (mean = –0.03), although pedigree records note the occurrence of matings between related animals in corn cribs CC104, CC303, and CC603 and field cages NC6, NC10, NC12, NC16, and NC18. Corn cribs CC102 and CC602 (F_is = –1.00) and field cages NC17 and NC20 (F_is = –0.17 and –0.24, respectively) contain the most outbred breeding groups; all other corn cribs and field cages showed no evidence of nonrandom mating according to their F_is values. The percentage of hybrids animals within field cages had no apparent influence on F_is values.

Genetic diversity.

The numbers of actual and effective STR alleles and the observed (OH) and expected (EH) heterozygosity estimates among the indoor colony (Table 2), corn cribs (Table 3), and field cages (Table 4) reveal that although overall OH and EH values in the indoor and outdoor colonies were comparable, the indoor colony and field cages exhibited slightly more alleles than did corn cribs, probably because of the smaller sample sizes from corn cribs. Whereas 17.4 actual alleles per locus were observed overall in outdoor field cages (Table 4) the average numbers of alleles per group ranged from 8.2 to 13.8, reflecting a deficit of between 3 and 9 alleles per group, due to genetic subdivision, nonrandom mating, and genetic drift.

Table 2.

Estimates of actual (na) and effective (ne) allele numbers, observed (OH) and expected (EH) heterozygosity values, and pairwise F_stvalues among the different housing areas in the indoor colony

n		AB	AW	BB	EX	HO	MB	NW	QU	SS	SW	THPAD	Overall/mean
	na	13.60	14.00	10.60	11.80	12.53	7.67	13.00	6.27	16.47	3.21	8.33	17.87/11.28
	ne	5.56	5.55	5.06	5.20	5.36	4.86	5.44	4.28	5.71	2.72	4.43	5.58/4.8
	OH	0.74	0.76	0.77	0.72	0.74	0.74	0.77	0.74	0.74	0.64	0.70	0.74/0.73
	EH	0.77	0.77	0.77	0.76	0.77	0.77	0.79	0.78	0.77	0.69	0.73	0.77/0.76
134	AB
148	AW	0.00
54	BB	0.00	0.00
62	EX	0.00	0.00	0.00
100	HO	0.00	0.00	0.00	0.00
13	MB	0.00	0.00	0.00	0.00	0.00
108	NW	0.00	0.00	0.00	0.00	0.00	0.00
8	QU	0.00	0.01	0.01	0.00	0.00	0.00	0.00
381	SS	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
3	SW	0.02	0.02	0.03	0.02	0.02	0.01	0.03	0.00	0.02
21	THPAD	0.00	0.01	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.03

AB, modular buildings; AW, animal wing; BB, Butler Building; EX, exposure facility; HO, hospitals; MB, neurobehavioral testing rooms; NW, North Wing; QU, quarantine rooms; SS, Speed Space; SW, South Wing

Table 3.

Estimates of actual (na) and effective (ne) allele numbers, observed (OH) and expected (EH) heterozygosity values, and pairwise Fst among the different corn cribs

Table 4.

Estimates of actual (na) and effective (ne) allele numbers, observed (OH) and expected (EH) heterozygosity values, and pairwise F_stvalues among the different field cages (NC)

The configurations of the corn cribs (that is, whether single male–multiple female caging or multimale–multifemale caging) did not appear to affect genetic diversity estimates according to observed and expected heterozygosity values. In agreement with a previous study,¹⁹ the pure-Chinese animals in NC17 were the most genetically diverse, in light of their high average observed and expected allelic numbers and heterozygosity estimates (Table 4). In comparison, field cages with only pure-Indian animals (NC2, NC5, and NC15) exhibited much lower-than-average estimates of observed heterozygosity (0.77) and expected heterozygosity (0.74). The proportion of hybrid animals in field cages did not appear to strongly influence the genetic diversity within any of the field cages. For example, field cages NC3 (n = 97), NC12 (n = 166), NC19 (n = 25), and NC20 (n = 75), for which 0.10, 0.14, 0.09, and 0.13 of the populations, respectively, were hybrids, exhibited comparable observed and expected heterozygosity estimates (0.75 or more) to NC14 (n = 82) and NC18 (n = 129), for which 0.47 and 0.71 of the populations, respectively, were hybrids (Table 4).

Genetic differentiation.

Varying genetic differentiation (Tables 2 through 4) is present among the different housing facilities at the CNPRC. Low pairwise F_st estimates are indicative of overlaps in allele frequencies between the populations being compared. These estimates reveal that almost all housing areas within the indoor colony, except the South Wing and temporary housing pad, contain animals that are not genetically differentiated. The South Wing contains the most genetically discrete animals of all groups studied, probably due to sampling effects, because only 3 animals from the South Wing were used in this analysis. The animals in quarantine rooms were most divergent from animals in the animal wing and Butler Building. Among corn cribs, variable degrees of genetic differentiation were observed that ranged from no differentiation (–0.14 between CC602 and CC807) to high levels of differentiation (0.64 between CC102 and CC602), but most, if not all, of these calculations probably reflect biases due to the small samples of animals from many of the corn cribs. In general, estimates of genetic differentiation were highest among the SPF and superSPF macaques (Table 4) as well as being most relevant, because these populations represent derived, well-established and closed breeding groups. Pairwise F_st computations also revealed moderate genetic differentiation (0.05 to 0.11) between NC17, the field cage with pure-Chinese animals, and all other field cages, including those with high proportions of Chinese–Indian hybrids.

Discussion

Random-bred populations of nonhuman primates are deemed suitable for studies of human diseases because random breeding preserves genetic variability in large populations. These populations can contain mutants that mimic disease processes or subjects that exhibit specific response characteristics that cause them to be susceptible to disease agents. Although colony-bred animals usually are assumed to be random-bred, closer examination of these populations over time and across centers has revealed evidence of founder effects and other genetic restrictions, including effects from nonrandom mating and genetic drift that have promoted loss of genetic diversity, occurrence of genetic subdivision, and increases in inbreeding coefficients over time.^19,20,21,22 Previous studies^19,30 have demonstrated that founder effects, intergenerational drift, nonrandom mating, and genetic bottlenecks affect allelic diversity more profoundly than they alter heterozygosity, due to the disproportionate loss of rare (compared with common) alleles. Therefore, although the average levels of expected heterozygosity in the 20 outdoor field cages of macaques (Table 4) ranged from only 0.69 to 0.87, the average number of alleles per locus ranged from 8.2 to 13.8.

The ancestry of rhesus macaques must be verifiable when animals with defined genetic or phenotypic characteristics are desired for research or breeding management purposes. The inadvertent inclusion of rhesus macaques from different geographic origins, such as China and India, or their hybrids, in experimental groups can foster results that are ambiguous because the elevated genetic contribution to phenotypic variance can obscure or swamp the contribution of experimental treatment effects to the total phenotypic variance. These taxa, particularly those of full-Chinese and full-Indian origins, can exhibit profound phenotypic differences in physiologic and behavioral traits that are controlled by genetic mechanisms.^{3,8-11,17,19,26,31,33} Primates of different geographic affiliations, such as India and China, have been organized into discrete subspecies based on their morphologic differences,¹⁵ but whether or not these differences correlated with genetic differences such as those we studied here is unclear.

Despite carrying high overall levels of genetic (observed and expected heterozygosity of 0.75 or greater) and allelic (15 to 17 alleles; data not shown) diversity colonywide, the breeding groups or social units in the corn cribs and field cages at the CNPRC represent limited gene pools, and maintaining gene flow among them limits differentiation due to genetic drift and consanguineous mating. Concordant with a previous study,¹⁹ assessments here show that the field cage with pure-Chinese rhesus macaques was more genetically diverse than were cages with only Indian animals. Excessive inbreeding has been avoided with the development of accurate marker-based pedigree records and sufficient numbers of unrelated animals available for breeding. Balancing the sizes of matrilines or the selective harvesting of female macaques without undue risk of social disruption also has ensured that none of the matrilines becomes over-represented in the colony's breeding population. Similarly, the strategy of exchanging 1-wk-old infants between dams housed in separate field cages and corn cribs³² promotes genetic diversity and fosters effective gene flow that homogenizes the gene pools of different breeding groups. The ad hoc pairing of infants needing maternal care with receptive adult female macaques represents another mode for gene flow among the different housing facilities. The lack of genetic differentiation among the indoor housing areas, corn cribs, and field cages but marked genetic subdivision within each of these 3 breeding areas is consistent with expectations based on population genetics theory. Furthermore, the higher frequency of infant exchanges, in addition to their higher adult sex ratios and population size, has resulted in higher genetic diversity and lower genetic subdivision among the field cages compared with the corn cribs. While further increasing the adult sex ratio of all breeding groups would probably increase their genetic diversity, the strong correlation between male rank and reproductive success might mitigate this influence.

The field cages and corn cribs are used primarily as sources for animals for assignment to specific research projects. In some of the indoor locations, such as the hospitals and exposure facilities, macaques are kept on a very short-term basis, and most of these animals are not used for breeding and therefore do not affect the colony gene pool. However, many of the offspring that result from time-mated breeding indoors are moved outdoors as infants and integrated into the infant exchange and adoption programs. As such, animal movement between the indoor and outdoor housing facilities is bidirectional.

Between 2004 and 2008, between 1147 and 1259 animals annually were moved indoors for research on infectious and respiratory diseases, reproductive science, and regenerative medicine. These estimates reflect a 10% growth in animal usage during that time period and underscore an increasing trend in the demand for research stocks. In the present study, estimates of migration rates based on Wright method³⁷ (that is, Nm = (1 – F_st) / 4F_st) reveal that each year, about 420 field-cage and 250 corn-crib animals that were involved in our study would have been moved indoors. Given that only 52% of the current indoor animals were sampled here, these estimates are roughly consistent with the 2004–2009 annual average of 1200 animals translocated into the indoor colony and confirm the accuracy of our F_st estimates and Wright³⁷ model for estimating migration rates as colony management tools. Animal movement between outdoor and indoor colonies for research purposes and migration between field cages and corn cribs for social reasons promote gene flow and preserve genetic heterogeneity, so that populations housed in the 3 different facilities remain genetically diverse and relatively panmictic.

Calculations of F_is for the current indoor colony, corn crib, and field cages demonstrate that current management practices, including swapping infants and balancing matrilines, have been very effective in limiting inbreeding, shared kinship, and population substructure. By minimizing inbreeding within the field cages and corn cribs, the infant swaps also have proven to be a cost-effective way to increase the effective population size to a value closer to population size in each enclosure.^21,23 However, despite these efforts, as the colony experiences natural population growth and expansion, the rhesus macaque enclosures will contain limited numbers of half-siblings, full-siblings, and parent–offspring combinations, increasing the risk of consanguineous breeding. According to the pedigree records used here, a few matings between close relatives in several outdoor enclosures including corn cribs and field cages, although not reflected in the outdoor F_is estimates, have occurred. The bias toward higher F_is values in the modular buildings, animal wing, exposure facility, hospitals, North Wing, Speed Space, and temporary housing pad indoor housing areas may have been an artifact of researchers selecting genetically homogenous, related animals or those representing particular pedigrees for experimental purposes including inbred animals from the outdoor colony. Although animals with Mamu A*01 or B*01 alleles are selected as SPF colony breeders, this practice has not resulted in excess homozygosity in the SPF population or affected the overall genetic heterogeneity of the conventional or SPF populations.²⁰

A previous report showed that the conventional rhesus stock at the CNPRC exhibits intergenerational genetic drift, probably due to the varied housing conditions of the nonSPF animals (including outdoor field cages, corn cribs, and indoor housing units) and the size and composition of their multiple breeding groups.²⁰ The considerable genetic divergence (high F_st values) among corn cribs and field cages in the current study might have been expected to have been accompanied by high F_is values and an indication of increased matings between close kin within breeding groups in these enclosures. Although infant cross-fostering and the selective removal of matrilines continue to minimize inbreeding effectively, the genetic differentiation that persists among the corn cribs and field cages indicates that the number of migrants that contribute to their new gene pools is too small to prevent allele fixation or extinction. Although one (reproductively successful) migrant per generation is sufficient to mitigate the effects of genetic drift and prevent particular alleles from being lost under equilibrium conditions (that is, no selection, mutation, migration, or inbreeding),¹⁶ the CNPRC rhesus colony does not simulate such an ideal population, probably due to nonrandom mating within each breeding group. Therefore, even greater gene flow should be promoted to forestall genetic drift and allele loss from the gene pools of the corn cribs and field cages. For instance, because they are cost-effective and easily implemented, the infant-exchange and infant-adoption program should be expanded to include more infant migrants between enclosures per breeding season, to rigorously counter the effects of genetic drift fostered by nonrandom mating. However, the SPF status of the macaques constrains the cages among which animals may be moved.

Other management options, including the periodic replacement of sexually mature male macaques by unrelated males from other field cages and corn cribs (a method that might be difficult to implement due to strict rhesus social structures), culling male breeders after siring a designated number of offspring to allow younger breeders opportunities for reproduction (that is, balancing patrilines), and increasing the sex ratio (numbers of male per female monkeys) of breeders in the field cages, could reduce genetic subdivision among outdoor enclosures and decrease overall inbreeding levels outdoors.²³ However, social stability is positively correlated with sex ratios that are slightly biased toward females (that is, more female than male monkeys).²³ One other consideration proposed previously¹³ is to curb excessive losses in allelic diversity by preferentially choosing male juveniles for harvest as future breeders according to their possession of rare alleles at loci genotyped for parentage testing and population genetics. This strategy would require sampling a sufficiently large number of loci that are representative of the diversity of their entire genomes and would only be practical using high throughput genotyping of single nucleotide polymorphisms. Another approach for maintaining rare alleles and genetic diversity is the exchange of males among national primate research centers. This option has become more feasible with the standardization of genetic markers and federation of genetic databases through the work of the National Nonhuman Primate Consortia.^18,24

Currently, the long-term breeding populations in the outdoor field cages at the CNPRC include rhesus macaques of Indian and Chinese origin as well as a very small minority of animals with Vietnamese and Burmese origins. Although some groups of pure-Chinese and pure-Indian rhesus macaques now are maintained separately, historically this was not the case, and widespread dispersal of Chinese rhesus genes throughout the CNPRC colony has occurred. Owing to the high dispersal rate among the colony animals, random animals of various degrees of admixture are distributed throughout the colony.¹⁹ The prevalence of animals of mixed Chinese–Indian ancestry in the CNPRC rhesus colony revealed by the STRUCTURE analysis is an ironic testament to the successful effort by colony management to distribute genetic diversity by encouraging gene flow (Figure 1).

Even though the CNPRC SPF colony is used for propagation, the conventional nonSPF colony forms the basis of the SPF animal supply. As a consequence of their method of derivation, the SPF animals at the CNPRC, such as those in field cages NC19 and NC20, are not genetically distinct from the rest of the field cages, including NC17. These results, together with previous findings²⁰ demonstrate that the CNPRC SPF colony, including the Indian and Chinese superSPF subpopulations, consist of a large number of genetically diverse animals with minimal genetic divergence from their conventional nonSPF founders. As reported previously,²⁰ the derivation of regular SPF and superSPF macaques directly from the conventional population forms a barrier to direct gene flow among the SPF and superSPF populations and limits gene flow among the different SPF populations (Table 4).

Single-cage housing in the indoor colony, outdoor corn cribs, and half-acre field cages represent the 3 primary types of enclosures at the CNPRC, the numbers of which vary depending on research needs, social groups and their size, and long-term breeding requirements. This study revealed that the genetic management policies at the CNPRC are focused on maximizing genetic heterogeneity within and among breeding groups by introducing geographically representative, genetically diverse, unrelated founder animals (as evidenced among the superSPF animals in NC17, which are first- and second-generation descendents of Chinese founders whose collective F_is value was –0.17). To that end, genetic management strategies at the CNPRC are designed to minimize genetic drift and genetic bottlenecks by promoting gene flow among the colony's housing facilities.

Despite management activities, including marker-based parentage testing, designed to prevent mating between close kin, some consanguineous inbreeding has occurred. The lack of population structure among the 3 separate housing facilities clearly demonstrates high dispersal of macaques among these facilities. However, significant effects of genetic drift, including loss of allele diversity, are present among the units within these housing facilities. Although comparable multigenerational STR data from other regional primate centers in the United States are unavailable, those centers likely are experiencing similar genetic restrictions and consequences, such as allele loss and genetic subdivisions, as expected in closed colony-bred populations like the CNPRC colony. To prevent further loss of rare alleles due to drift and inbreeding that can preclude the identification of allele–disease associations, more efforts are required to promote gene flow within the different types of housing facilities. The introduction of Chinese rhesus macaques to the colony in the early 1980s, in conjunction with management-mediated gene flow, has had a long-lasting effect on the colony's genetic composition. Being of intermediate genetic composition, hybrid animals respond differently to experimentation than do either unmixed parental (full-Indian or full-Chinese) population, and any inadvertent inclusion of Chinese–Indian hybrid animals in experimentation for which purebred animals are desired could obscure treatment effects.