A Comparative Analysis of the Genetic Structures of Rhesus Macaques (Macaca mulatta) Kept in Enclosure-Based Systems and Semi-Natural Settings

Abstract

This study compares 2 rhesus macaque populations, one captive (the California National Primate Research Center [CNPRC]) and one seminatural (the Cayo Santiago [CS] population), to evaluate the impact of colony management on genetic structure. While the CNPRC houses its 4,000-plus animals in different housing enclosures, CS has a seminatural, free-ranging population of ∼1,900 individuals. Pedigrees of 3,679 animals at the CNPRC and 2,147 animals at CS were analyzed to assess cross-generational genome uniqueness (GU), inbreeding coefficients, mean kinship (MK), founder equivalents, and founder genome equivalents. An estimated 30% of the 409 CS founders contributed to the present population, compared with <10% of the 2,980 founders and subsequently introduced animals at the CNPRC. Differences in pedigree structure and potential inbreeding risk, particularly at the CNPRC, indicate a more genetically subdivided population there than at CS. The CNPRC also exhibited a lower GU than CS, whereas the MK of both colonies showed a similar upward trend over the study period (2000-2024). Also revealed are significant differences in reproductive activity, lifespan, and the number of families per individual between the 2 colonies. There are challenges in maintaining captive-bred colonies, such as the CNPRC, where restricted gene flow and inbreeding lead to dramatic declines in genetic variability. Strategies such as equalizing reproductive success, implementing cross-fostering programs for infants, and introducing new animals are insufficient to combat inbreeding, genetic subdivisions, and the loss of founder representation. Findings highlight the need for ongoing monitoring of inbreeding coefficients, GU, and MK to safeguard genetic integrity and ensure the long-term viability of captive-bred colonies. The analysis also offers insights for managing these colonies, highlighting MK as a key metric for preserving genetic diversity because it integrates founder representation, drift, and inbreeding into a colony-wide measure that directly informs colony management decisions.

Introduction

The National Center for Primate Biology, the precursor to the CNPRC, was designed to house a foundation stock of 400 animals from the Indian subcontinent to establish the CNPRC colony with the infrastructure to support up to 1,000 animals eventually.¹ Over time, however, the population expanded significantly, reaching >4,000 animals. The primary mission of the CNPRC’s captive rhesus macaque (Macaca mulatta) colony is to support biomedical research initiatives such as infectious diseases such as AIDS, vaccine development, xenotransplantation, reproductive sciences, neurobiology (including cognitive function and neurodegenerative conditions such as Alzheimer disease), respiratory disorders such as asthma, biobehavioral organization, behavioral development, social relationships, eye disease, and primate physiology, including therapies for chronic colitis, nutritional deficiencies, reproductive biology, and lysosomal storage disease.^2–21

Rhesus macaques in the CNPRC colony are housed in single-cage indoor housing, half-acre outdoor field cages, and corncribs. There are 23 half-acre field cages (∼2,270 animals), with each accommodating up to 184 animals. The tenure of these field cage populations as a breeding resource varies in length from 1 year (recent field cage formations) to 20 years. The breeding system in field cages is multimale, multifemale, with an average sex ratio of 3 females per male. The annual live birth rate in the field cages, calculated as the number of live births divided by the total number of females aged ≥3.5 years plus younger females that gave birth during the same year, is 78%, with an average yearly production of 700-800 newborns. As the number of field cage occupants exceeds its optimal level, surplus animals are used to populate new production field cages. Animals are also relocated between field cages for social management (addressing issues such as deleterious aggression, matriline fragmentation, sex ratio balance, and group stability) and to optimize gene flow within field cages and corncribs as well as among these enclosures.^22–24 There are 36 corncribs (∼413 animals), each accommodating up to 27 animals, and providing options for smaller social groups, including young, weaned animals in peer-housing situations with an adult animal, and harem breeding groups that comprise one male and multiple females in the reproductive stage who will typically experience their first birth. The single-cage indoor rhesus population comprises approximately half of the colony animals (∼1,970 animals), designated for various research projects, including time-mated breeding for reproductive studies and those requiring level 2 and level 3 biosafety containment for infectious disease research. This indoor colony also encompasses nursery-reared, weanling, and aged animals (>18 years). While the annual proportion of animals allocated to terminal studies and short-term projects has remained relatively stable, the makeup of animals in certain indoor enclosures, such as hospitals and exposure facilities, is more dynamic and changes on a daily, weekly, and monthly basis.

Due to its enclosure-based housing system, the CNPRC operates as a genetically subdivided colony, with distinct genetic differences among enclosures, indicating that the colony population is not uniformly genetically homogeneous.^22,24,25 As a result, these captive-bred individuals may exhibit increased variability in their responses to biomedical experimentation.²⁶ Consequently, conducting these experiments may require more animals than are currently available.²⁷

The CNPRC’s rhesus macaque colony can also be considered a semi-closed population; that is, it was established from multiple external sources but subsequently maintained primarily through internal breeding, with only limited immigration to supplement genetic diversity. The scarcity of new introductions, such as wild-caught animals, combined with inadequate management of these resources, leads to inbreeding within captive populations.^28–30 Therefore, the semi-closed colony configuration of breeding rhesus macaques can lead to a significant decline in genetic diversity over time due to the random loss of alleles as a result of genetic drift and inbreeding, which can also impact the genetic variability underlying important biomedical traits, complicating the interpretation of results from these experiments when applied to human disease.^31,32 Over time, these animals can become prone to genetic diseases and inbreeding depression, resulting in fewer viable offspring and increased morbidity and mortality.^27–30

The acknowledgment of the challenges of captive breeding of rhesus macaques has led the colony managers at the CNPRC to implement genetic management strategies to conserve genetic diversity and enhance the production of genetically characterized animals.^33–37 Preserving genetic diversity in captive rhesus macaques is essential for both colony sustainability and the reliability of biomedical research. Reduced diversity increases inbreeding depression, lowers fertility, and heightens disease susceptibility, while diverse populations provide more robust and translational models for human disease.^27–29,38 Therefore, effective long-term colony management strategies require constant adaptation and implementation of genetic management goals aimed at maintaining genetic diversity in response to current needs. Given the decades-long genetic management efforts at the CNPRC, it is essential to assess the program’s efforts in maintaining genetic diversity and minimizing inbreeding and genetic subdivision.^{23,32–37,39,40}

To assess the effectiveness of the CNPRC’s genetic management efforts, primarily their impact on genetic diversity, this study aimed to comparatively evaluate the population genetic diversity and structure of the CNPRC against those of another rhesus macaque colony, namely the CS colony, managed by the Caribbean Primate Research Center. CS is an island that measures 0.14 km² and is located ∼1 km off the coast of Puerto Rico.

The rhesus macaque colony on CS was established in December 1938 as a seminatural population with 409 initial colony members imported from India. Since its inception, the colony has remained closed, with no animals added, thereby preserving its genetic continuity and making it a uniquely stable population for long-term research.^41–43 The CS rhesus macaque colony is considered a feral macaque population because it is not native to Puerto Rico, but it consists of a free-ranging population that exhibits species-typical behaviors, such as grooming, mating, and dominance hierarchies.^44–47 Similar to the CNPRC animals, CS animals receive food ad libitum and live without natural predators, but with minimal veterinary care and significant ecological pressures.⁴⁸ The CS population comprises ∼1,900 individuals (∼900 males and ∼990 females), which (at the time of writing) can be subdivided into 11 naturally formed multimale and multifemale social groups and one all-male group, with 5 adult individuals. While there are social groupings among its colony animals, the CS animals are not enclosed or genetically subdivided, as they freely roam between social groups, especially young males.⁴⁹

This free migration contributes to greater genetic cohesion across the CS colony, which in turn influences its effective breeding population size (N_e), a key metric in population genetics.⁵⁰ N_e refers to the number of breeding individuals who contribute offspring to the next generation and is closely tied to the size and social structure of a colony. Because of its free-ranging, socially integrated design, the N_e relative to actual population size at CS is expected to be substantially higher than that of the CNPRC. This distinction is critical: a smaller N_e increases the likelihood of shared ancestry and accelerates the accumulation of homozygosity across generations, thereby reducing genetic diversity and elevating the risk of inbreeding depression.⁵¹ In contrast, a larger N_e helps preserve genetic variation and slows the rate of inbreeding, making it essential for sustaining genetic health in both conservation and biomedical research contexts.⁵¹

Similar to the CNPRC, the CS colony serves as a critical resource, primarily for noninvasively studying various aspects of a free-ranging rhesus macaque population such as primate behavior, demography, social structures, and genetic evolution in an environment that is both controlled and semi-wild, thereby providing crucial insights into the natural behaviors that form the foundation of primate and human biomedical and behavioral studies.^{41,42,52–56}

The live birth rate on the island is slightly >82%, averaging 322 newborns per year over the past 10 years. Mainly in response to the need to stabilize population growth on the island, animals are removed from CS and either transferred for sale, incorporated into breeding colonies at the Sabana Seca Field Station, or leveraged for research.^49,52 This research encompasses the development and testing of vaccines, microbicides, preexposure prophylaxis methods, and advanced HIV therapies.^16,57–59

Similar to the CNPRC, the CS colony has benefited from genetic management strategies grounded in a DNA-based genealogy established through systematic parentage testing in the 1990s.^{23,32–34,39,40,49,60} Comparable housing systems, including field cages, corncribs, and single housing, are also used across at least 6 other U.S. National Primate Centers (Oregon, Southwest, Tulane, Washington, Wisconsin, and Emory NPRCs). The availability of consistent and concurrent DNA datasets and parentage and pedigree analytical methods for both the CNPRC and CS populations, despite differences in housing system approaches and rearing conditions (captive compared with seminatural or semi-wild), provides a valuable foundation for the comparative analyses presented herein.^{23,25,32,39,40}

As such, the CS population serves as a suitable comparative group or natural (wild) animal representatives for comparing the effects of population-level changes in the diversity and structure of the CNPRC study animals. Therefore, by comparing the population genetics of the CNPRC and CS populations and identifying deviations from natural conditions at the CNPRC, more effective (and natural) strategies for managing and preserving genetic diversity may be developed, ensuring the health and sustainability of captive populations such as the CNPRC colony.

This study examines the multigenerational pedigrees from the animals that are presently being maintained in breeding (field cage) and holding (corncrib) facilities at the CNPRC and the free-ranging (semi-natural) CS population, comparing key parameters such as inbreeding and kinship, and founder genome equivalents (Fg).^31,61

Materials and Methods

Animals involved in this study were managed in accordance with IACUC expectations and NIH guidelines, which prescribe humane care and use of laboratory animals. This study adheres to the protocols approved by the IACUCs at the University of California and the University of Puerto Rico, as well as the American Society of Primatologists “Principles for the Ethical Treatment of Nonhuman Primates.” All research reported in this study was conducted following the legal requirements of Puerto Rico.

This study used the PEDSYS program to construct and analyze the CNPRC’s and CS’ rhesus macaque pedigrees.⁶² Genealogical and demographic records from these pedigrees were used to calculate genome uniqueness (GU), inbreeding (F), mean kinship (MK), founder equivalents (Fe), and Fg for living animals in each colony, after excluding individuals no longer present in the population (ie, deceased or removed), using PEDSYS.^62,63

A total of 400 individuals founded the CNPRC colony, and 409 founded the CS colony.^1,43 However, not all members of this foundation stock contributed genetically to the present-day populations. For this study, ‘founders’ refer to those whose genetic material has been transmitted to the extant populations, and their genetic composition forms the ancestral basis from which all subsequent generations in these colonies have descended.

GU was estimated using a gene-drop simulation, which assigns distinct alleles to each founder in the current pedigree structure and models their transmission through successive generations according to Mendelian inheritance.⁶⁴ The uniqueness value for an individual is calculated as the proportion of simulations in which that individual inherits the sole copy of a founder allele.²⁹ Therefore, this metric reflects the probability that an animal carries founder alleles that are rare and potentially at risk of being lost from the population, and prioritizing individuals with the highest GU, rather than focusing solely on minimizing co-ancestry or kinship, ensures that the preservation of these rare genotypes takes precedence.²⁶

F is the probability that an individual inherited an identical homologous allele from a common ancestor, and MK is an individual’s average kinship with all other individuals living in the colony and reflects the expected proportion of genomic similarity stemming from identity by descent.^21,28 Therefore, MK values reflect the degree to which an individual is related to other animals within the colony. In Table 1, MK is calculated for all living animals in the study sample over the entire study period, calculated using the complete colony-wide pedigrees. In contrast, Table 2 presents MK values for living study animals, grouped into 5 successive birth cohorts, to illustrate age-stratified, generational differences in kinship estimates across cohorts. MK is crucial in breeding programs designed to preserve genetic diversity in subdivided and/or small populations, such as social groups, breeding cages, and matrilines. For example, when selecting between 2 potential sires for mating, and disregarding other considerations, it is preferable to choose the sire with a lower MK value.

Table 1.

Pedigree-Based Estimates of the Actual Number of Founders, Founder Equivalents, Founder Genome Equivalents, Mean Kinship, and Genome Uniqueness Averaged Across All Study Animals at the CNPRC and CS That Were Living During the Period of Study (2000-2024)

Population and genetic metrics	CNPRC	CS
No. of animals included	3,679	2,147
Size of foundational stock	2,980	409
Fe	236.18	129.14
Fg	130.15	110.76
MK	0.50%	0.70%
GU	1.00%	4.00%
No. families for males	0-42 (6.77 ± 7.80)	0-45 (5.22 ± 6.00)
No. families for females	0-12 (2.52 ± 1.96)	0-14 (2.61 ± 2.53)
No. of animals included	14,600	9,300
Lifespan of males	7-35a (7 ± 2) y	7-27a (8 ± 2) y
Lifespan of females	6-36a (7 ± 2) y	6-29a (9 ± 3) y
Reproductive span of males	3-26 (7 ± 2.5) y	3-25 (8 ± 2) y
Reproductive span of females	2-24 (7 ± 2.5) y	2-21 (9 ± 3) y
No. generations	12	10

Also presented are the numbers of families across males and females. Numbers in parentheses are averages and SDs (±) for each estimate. Estimates of lifespan include all causes of death, including those assigned to terminal studies.

Fe, founder equivalents; Fg, founder genome equivalents; GU, genome uniqueness; MK, mean kinship.

^a Maximum lifespan.

Table 2.

Pedigree-Based Average Estimates of Inbreeding Coefficient, Genome Uniqueness, and Mean Kinship at the CNPRC and in CS Over 5-y Intervals From 2000 to 2024

	N		F		GU		MK
	CNPRC	CS	CNPRC	CS	CNPRC	CS	CNPRC	CS
2000-2004	96	97	0.009 ± 0.033	0.0029 ± 0.012	0.105 ± 0.042	0.087 ± 0.193	0.002 ± 0.0005	0.005 ± 0.0002
2005-2009	238	306	0.015 ± 0.033	0.0042 ± 0.015	0.046 ± 0.032	0.082 ± 0.151	0.004 ± 0.0005	0.005 ± 0.0005
2010-2014	607	536	0.013 ± 0.039	0.0046 ± 0.015	0.017 ± 0.021	0.04 ± 0.103	0.005 ± 0.0005	0.006 ± 0.0005
2015-2019	1,058	724	0.020 ± 0.041	0.0046 ± 0.012	0.005 ± 0.020	0.037 ± 0.101	0.005 ± 0.0005	0.007 ± 0.0005
2020-2024	1,680	484	0.026 ± 0.039	0.0070 ± 0.012	0.002 ± 0.020	0.009 ± 0.061	0.006 ± 0.0005	0.009 ± 0.0005
Mean	735.8	429.4	0.018 ± 0.037	0.0050 ± 0.013	0.009 ± 0.027	0.038 ± 0.122	0.005 ± 0.0005	0.007 ± 0.0005

N denotes the number of living individuals born during the specified period (birth cohort). Genetic metrics (F, GU, MK) for these individuals were calculated using the colony-wide pedigrees and represent age-stratified differences.

F, inbreeding coefficient; GU, genome uniqueness; MK, mean kinship.

Lacy’s methods were used to generate colony-wide estimates of Fe and Fg.⁶⁵ The values of Fe and Fg reflect the theoretical average number of equally contributing founders, which would be expected to cause the present level of genetic diversity.⁶⁵ Fe values were calculated by assigning unique alleles to each founder and simulating their transmission through the pedigree using Mendelian inheritance.⁶⁵ Fe values assume that no effects from genetic drift have influenced colony-wide allele frequency, while the computation of Fg assumes otherwise. This is because Fe values are calculated from the squared proportional genetic contributions of each founder, reflecting the unequal representation of these founders in the current population. The estimation of Fg values, however, refines this by simulating allele transmission using the gene-drop analysis through the pedigree, accounting for both unequal contributions and genetic drift over generations. Therefore, an observation that Fe is greater than Fg indicates the presence of significant genetic drift affecting the estimated number of founders.

The lack of animals with complete genetic and parentage data prior to 2000 prevented their inclusion in the pedigree analysis. An additional 28 Caribbean Primate Research Center animals (0.76%) that lacked parental data, likely reflecting the feral nature of the population where testing all potential parents was not feasible, were not included in subsequent analyses. When parentage information is missing in pedigree analysis, F is typically underestimated because gaps in ancestry obscure shared relationships, making true inbreeding appear lower than it is. GU, on the other hand, is overestimated because unknown parents are often treated as unrelated founders, thereby falsely inflating perceived diversity.

Similarly, MK is typically underestimated because incomplete pedigrees reduce calculated relatedness among individuals. Fe and Fg are generally both inflated, as each missing parent is assumed to be a unique founder contributing novel genetic material, which misrepresents the actual genetic structure. Overall, missing parentage creates an illusion of greater diversity and lower relatedness than truly exists. Therefore, to minimize bias in metric estimation, subsequent population genetic analyses included only living animals with complete genetic parentage data during the 5-year intervals of the study period (2000-2024), comprising 3,679 individuals from the CNPRC and 2,147 from CS. All statistical analyses were performed to compare observed findings against expected background variation, with significance defined at the accepted threshold of P < 0.05.

Results

Pedigree records from ∼7,200 male and ∼7,400 female (N = 14,600) CNPRC individuals and ∼4,200 male and ∼5,100 female (N = 9,300) CS animals were analyzed. The CNPRC records represented 12 generations of pedigree data, whereas the CS records represented 10 generations of pedigree data. As MK, GU, Fe, and Fg are relative measures, they were calculated for all living animals (ie, after trimming out the dead and removed individuals) with complete parentage data, 3,679 at the CNPRC and 2,147 at CS, over the 25-year study (Table 1). Table 1 also presents demographic information for these animals. In addition, MK, GU, Fe, Fg, and F values were computed for each 5-year interval within the study period (Table 2). Due to the extended time frame the study animals represented, and the transfer of animals among indoor and outdoor enclosures, it was not possible to separate the CNPRC pedigree data for the field cages and the corncribs.

Based on these records, the maximum lifespans of males and females in the CS colony are 27 and 29 years, respectively. In contrast, at the CNPRC, males and females have maximum lifespans of 35 and 36 years, respectively (Table 1). Based on the maximum lifespan of animals in the 2 colonies, CNPRC animals tend to live longer than CS animals. The oldest CNPRC males and females alive at present are 19 and 18 years old, respectively. The oldest CS males and females are 21 and 19, respectively. At the CNPRC, males are reproductively active between the ages of 3 and 26 (7 ± 2.5) years, while females are reproductively active between the ages of 2 and 24 (7 ± 2.5) years. At CS, males aged 3-25 (8 ± 2) years and females aged 2-21 (9 ± 3) years are reproductively active (Table 1). In both populations, females become reproductively active earlier than males, but they also cease reproductive activity sooner. CNPRC males have between 0 and 42 families (mean = 6.77), where a family is defined as the set of offspring produced by a unique sire/dam pairing. Specifically, each family attributed to a sire comprises all individuals born from a single dam mated with that sire, representing a full-sibling group. CNPRC females have between 0 and 12 families (mean = 2.52) (Table 1). In CS, males have between 0 and 45 families (mean = 5.22), while females have between 0 and 15 families (mean = 2.61) (Table 1).

F ranged from 0.009 to 0.026 and averaged 0.018 in the CNPRC colony, and from 0.0029 to 0.007 and averaged 0.005 in the CS colony. Seventy-six percent (N = 2,783) of the 3,679 CNPRC animals surveyed exhibited an F value of 0%. The same fraction of the 2,147 CS animals, 76% (N = 1,643), exhibited F values of 0. Analysis of genomic similarity revealed that MK values in both colonies ranged between 0% and 1% (mean = 0.5% and 0.7%, respectively), while GU values ranged between 0% and 67% (mean = 1.00%) at the CNPRC and 0% and 88% (mean = 4.00%) in CS (Table 1).

Estimates of Fe were 236.18 for the CNPRC and 129.14 for CS, while Fg was 130.15 for the CNPRC and 110.76 for CS (Table 1). The CS estimates are consistent with earlier work by Kanthaswamy et al.⁴⁹ based on 1,227 CS animals or nearly 2-fold fewer than in the present study, which yielded Fe and Fg values of 53.83 and 38.07, approximately half the current estimates in founder metrics. The current estimates suggest that it would have taken 130-236 (CNPRC) and 111-129 (CS) equally contributing founders to achieve the current levels of F, GU, and MK at the CNPRC and in CS (Table 1). All 3 metrics are shaped by founder representation, either directly (as in GU and Fe, ie, Fe or Fg) or indirectly through its impact on genetic structure (as in F and MK). Furthermore, the relative discrepancy between Fe and Fg of 45% at the CNPRC and 14% in CS indicates the approximate loss of unique founder alleles due to genetic drift, resulting in a GU of 1.00% at the CNPRC and 4.00% in CS.

These data suggest the CNPRC experienced a greater loss of unique alleles, which undoubtedly resulted from the markedly greater genetic subdivision at the CNPRC compared with CS. Based on the findings above, F is determined mainly by N_e, rather than the total census population of a colony. This crucial distinction clarifies why F is significantly higher at the CNPRC compared with CS, as genetic subdivision has a substantial impact on N_e.

Table 2 presents F, GU, and MK estimates for living study animals grouped into successive 5-year birth cohorts, providing an age-stratified view of how kinship and genetic contributions vary across cohorts within each population. N denotes the number of animals born within each interval who remain alive and were included in the pedigree analyses. Based on N (total population size), the CNPRC exhibits a more pronounced increase, from N = 96 (2000-2004) to 1,680 (2020-2024), indicating consistent population growth. In contrast, the CS population initially grew steadily from 2000-2004 (N = 97) to 2015-2019 (N = 724) but declined in size by nearly half (N = 484) by the end of the study period in 2020-2024, because of the removal or culling of animals with the lowest GU to control population growth on the island. The combined F values in both colonies range from 0.005 to 0.030. The F values for the CNPRC show a significant increase over time, starting from 0.009 in the 2000-2004 interval and steadily rising to 0.026 by the 2020-2024 interval. In contrast, the F values for CS have remained relatively stable, starting at ∼0.003 and increasing only marginally, to 0.007 by the 2020-2024 interval. Overall, the table highlights a notable increase in F values for the CNPRC over 2 decades, while the CS values have remained relatively constant. In addition, Table 2 displays the changes in GU values from 2000 to 2024 for both colonies, ranging from 0.002 to 0.105. Overall, the table indicates a more dramatic decrease in GU values over 2 decades at the CNPRC than at CS, reflecting a more substantial change in the genetic makeup at the CNPRC. Taken together, these metrics reveal how founder representation, genetic drift, and breeding structure shape the long-term genetic health of each colony, underscoring the need for proactive management to prevent diversity loss and inbreeding accumulation.

Changes in MK values over the same 5-year periods in both colonies range from 0.002 to 0.009 (Table 2). Annual MK values at the CNPRC started at 0.002 during the 2000-2004 period, steadily increased to 0.004 in the 2005-2009 period, then increased slightly to 0.005 in the 2010-2014 and 2015-2019 periods, and finally peaked at 0.006 in the 2020-2024 period. The value of MK at CS started slightly higher at 0.005 in the 2000-2004 period, increased to 0.006 in the 2010-2014 period, and then further increased slightly to 0.007 in the 2015-2019 period, before finally rising to 0.009 in the 2020-2024 period. Overall, Table 2 highlights the increasing trends in MK for both the CNPRC and CS over the specified periods.

An ANCOVA was used to compare GU, MK, and F across colonies, controlling for one variable while testing correlations with another. An ANCOVA of the CNPRC data revealed that, after controlling for GU, there was a statistically significant difference (P = 0.001) in MK between sexes, with sex accounting for 1.4% of the variance in MK, and a weak but significant negative correlation between GU and MK (r = −0.18; P < 0.001). However, when controlling for MK, there was no significant difference in GU between sexes, with sex accounting for 0.0% of the variance in GU, and a weak but significant negative correlation between MK and GU (r = −0.18; P = 0.001). In addition, after controlling for F, there was a slight but significant difference in GU between sexes (P = 0.049), with sex accounting for 0.1% of the variance in GU, holding F constant, and a weak but significant negative correlation between F and GU (r = −0.06; P = 0.048). When controlling for GU, there was a slight but significant difference in F between sexes (P = 0.008), with sex accounting for 0.2% of the variance in F, and a weak but significant negative correlation between GU and F (r = −0.06; P = 0.008).

Similarly, an ANCOVA analysis of the CS data showed that after controlling for GU, there was a significant difference in MK between sexes (P = 0.001), with sex accounting for 3.0% of the variance in MK, and a moderate but significant negative correlation between GU and MK (r = −0.41; P < 0.001). When controlling for MK, there was no significant difference in GU between sexes (P = 0.454), with sex accounting for 0.0% of the variance in GU, and a moderate but significant negative correlation between MK and GU (r = −0.41; P = 0.001). Moreover, when controlling for GU, there was no significant difference in F between sexes, with sex accounting for 0.1% of the variance in F, and a weak negative but significant correlation between GU and F (r = −0.10; P = 0.001). After controlling for F, a significant difference in GU was observed between sexes (P = 0.001), with sex accounting for 0.8% of the variance in GU. A weak negative correlation was found between F and GU (r = −0.10; P = 0.001). No statistically significant difference in r was detected among F, GU, and MK values within either population. Furthermore, F, GU, and MK differ significantly between sexes in both populations.

The CS population represents a single, cohesive group (∼1,900 individuals) that is ∼9 to 238-fold larger than the largest of the social housing enclosures at the CNPRC, that is, the field cages and corncribs (Table 3). Consequently, the N_e at CS is significantly larger than at the subdivided CNPRC colony, calculated to be 824.39 individuals at CS, using N_e = 4N_mN_f/(N_m + N_f) (where N_m indicates males and N_f indicates females), as per Nei.⁵¹ The N_e across the field cages and corncribs at the CNPRC ranged from 0 to 57.25 individuals, indicating that the N_e at CS was at least 14-fold higher than that in field cage number 8, the enclosure with the highest N_e at the CNPRC. Because the CNPRC population also has a significantly smaller N_e than that at CS, genetic drift and inbreeding are more significant factors in the genetic makeup of the CNPRC colony. The results of Spearman correlation analysis indicated a significant negative association between F and N_e (r_s = −0.716, P = 0.000).

Table 3.

Population Size, Number of Adult Males and Females, Effective Population Size, and Average Inbreeding Coefficient Across the 2 Largest Types of Social Housing Enclosures at the CNPRC, Field Cages, and Corncribs

	Enclosure type	Population size	Nm	Nf	Ne	F
Field cage	1	140	10	41	32.16	0.005
	2	47	5	16	15.24	0.012
	3	81	10	22	27.50	0.009
	4	52	6	8	13.71	0.006
	5	205	16	52	48.94	0.004
	6	67	9	18	24.00	0.010
	7	49	5	2	5.71	0.014
	8	186	19	58	57.25	0.004
	9	66	1	8	3.56	0.004
	10	158	16	44	46.93	0.004
	11	29	0	12	0.00	0.013
	12	83	6	24	19.20	0.010
	13	129	5	49	18.15	0.006
	14	54	6	11	15.53	0.012
	15	206	17	62	53.37	0.004
	17	68	1	8	3.56	0.009
	18	99	11	27	31.26	0.008
	19	103	11	20	28.39	0.006
	20	172	17	55	51.94	0.004
	21	70	2	8	6.40	0.009
	22	194	18	55	54.25	0.003
	23	190	14	72	46.88	0.004
	24	28	3	0	0.00	0.011
Average		107.65	9.04	29.22	26.26	0.007
Corncribs	101	11	3	3	6.00	0.057
	102	8	0	8	0.00	0.057
	103	13	2	8	6.40	0.039
	303	21	1	2	2.67	0.030
	304	17	2	5	5.71	0.042
	402	21	3	2	4.80	0.033
	403	14	2	2	4.00	0.045
	404	11	2	5	5.71	0.063
	602	15	0	1	0.00	0.033
	703	9	2	4	5.33	0.078
	802	19	1	0	0.00	0.089
	805	18	2	2	4.00	0.028
	807	12	3	3	6.00	0.052
	900	15	1	5	3.33	0.048
	901	21	2	7	6.22	0.035
	902	16	2	0	0.00	0.039
	1,001	10	2	6	6.00	0.063
	1,003	16	1	0	0.00	0.089
Average		14.83	1.72	3.50	4.62	0.051

Data exclude enclosures designated for specific studies. The F values reflect the average estimates across all animals within each housing unit over the study period.

N_e, effective population size; N_f, number of adult females; N_m, number of adult males.

Discussion

The analysis of pedigree records from the CNPRC and CS colonies has provided valuable knowledge of their population genetic composition and structure. The study identified systematic differences in pedigree structure and potential inbreeding risks, especially at the CNPRC. The analysis also provides insight into each colony’s history, the living conditions of the animals, and their lifespan, all of which may have influenced their genetic diversity. The difference in population growth between the 2 populations, as reflected by the sample size, N, indicates that more than half the CNPRC animals included in this study represent the CNPRC’s long-term breeding resource. This resource has been optimized to maximize animal production, achieving a live-birth rate of 80% annually and an average of 700-900 newborns. On the other hand, given the need to stabilize its growth, the rhesus macaque population on CS is subject to routine selective removal of animals based on their sex and age.¹⁹ Furthermore, at CS, animals are not protected from harsh weather conditions such as hurricanes and other environmental hazards.⁴⁸

At the CNPRC colony, males and females demonstrated longer maximum lifespans than did males and females in the CS colony. This point highlights the comprehensive veterinary care strategies employed at the CNPRC to ensure the health and well-being of the animals while supporting various research initiatives. In contrast, the CS population, while genetically stable, lacks access to intensive veterinary interventions, which may contribute to its comparatively shorter lifespans. The CNPRC conducts comprehensive clinical measurements, at least annually, to evaluate the overall health of the animals. This includes complete hematology, serum chemistry, weight, and immunophenotyping by flow cytometry. The prevention of diseases and the ability to treat them contribute to longer lifespans. The CNPRC also employs a behavioral management strategy that includes social management, environmental enrichment, the human/animal interface, and environmental design.²⁴ Matriline-level cohesion and group demographics are regularly monitored to inform management decisions. These efforts enhance animal welfare and social stability.⁶⁶ The CNPRC animals also live in more controlled environments than in CS. A stable environment reduces the risk of stress, accidents, injuries, and even death.⁴⁸ There is also a higher likelihood of reduced competition for resources at the CNPRC, where the animals likely do not have to compete for food, water, or shelter, which reduces stress and allows them to thrive more effectively. However, captivity has been associated with elevated stress responses in rhesus macaques, evidenced by increased glucocorticoid levels and stereotypic behaviors, particularly under single housing conditions, although enrichment and social housing can mitigate these effects.^4,5,67–69

Building on these demographic differences, mating behavior differed between the 2 colonies. At the CNPRC, the animals remained reproductively active for a more extended period, with males averaging 1 year longer and females averaging 3 years longer than their counterparts in the CS colony. The longer reproductive span of the CNPRC animals compared with the CS animals corresponds with the observed differences in lifespan between animals from both populations. As reproductive success is an individual’s production of offspring over its lifetime, this supports the idea that reproductively successful animals at the CNPRC had better opportunities to continue contributing to the next generation of descendants.⁷⁰

The number of families also varied between the colonies. On average, the CNPRC males had slightly more families than did the males in CS, while the females in both populations had a similar number of families. These differences in the number of families may reflect variations in the reproductive strategies and living conditions of the animals, that is, being maintained in housing enclosures compared with being allowed to range freely. These captive conditions, that is, both the field cage (multi-male, multi-female system) and corncrib (harem system), promote proximity and enhance the reproductive capacity of CNPRC males, allowing them to monopolize access to females and father more offspring.⁷⁰ The rapid increase in the CNPRC colony’s population size described above is likely positively correlated with the number of families. The observed differences in male family numbers at the CNPRC compared with the CS could be attributed to the CNPRC’s practice of maintaining fewer adult males per breeding group. This suggests a more restricted male contribution to the overall gene pool at the CNPRC compared with CS.

These reproductive patterns directly influence genetic structure, as seen in the inbreeding coefficient (F) and genomic similarity (MK compared with GU values), which provided further insights into the genetic diversity and structure of the colonies. The colony management strategies, living conditions (captive compared with free-ranging), and the lifespan of these populations, as described above, could have influenced their genetic diversity and structure. Although the same percentage of CNPRC and CS animals (ie, 24%) exhibited an F value > 0, as noted in Table 2, the CNPRC rhesus macaques had a higher average F value (0.018) compared with the CS colony (0.005), which is most likely attributable to the housing conditions at the CNPRC.

In captivity and being confined to enclosures such as indoor housing, outdoor field cages, and corncribs, CNPRC animals face restrictions on movement that limit their ability to adjust their group membership and/or mate freely, leading to higher inbreeding rates. Infants may remain in the same group through sexual maturity, which can result in consanguinity, such as female offspring breeding with their sire. Conversely, the CS rhesus macaques generally have more opportunities to mate freely and avoid mating with relatives compared with those in captive settings.²⁴ Therefore, because these animals generally can avoid mating with close relatives, inbreeding in CS is found to be rare.⁷¹ The island’s 0.14-km² size offers ample opportunities for dispersal. This reduces the likelihood of mating with relatives compared with captive conditions, such as those at the CNPRC. Consistent with patterns observed in wild populations, the free-ranging rhesus macaques in CS actively avoid mating with close maternal kin.⁷¹ As in the wild, the sex-biased dispersal and reproductive strategies of the CS animals play a crucial role in determining the availability of kin, interactions among related individuals, and kin avoidance, which helps minimize inbreeding among maternal relatives. Wild females tend to remain in their natal groups and establish stable matrilineal hierarchies, while males leave their natal groups as they reach sexual maturity.^72–74 In captivity, as exemplified by the CNPRC colony, the extent to which rhesus macaques avoid inbreeding is directly proportional to their degree of relatedness to potential mates.^35,75 However, captive animals often have limited choices for mates, which leads to breeding with close paternal relatives, such as siblings or cousins.²⁴

The CS population represents a single, cohesive group (∼1,900 individuals) that is ∼9- to 238-fold larger than the social housing enclosures at the CNPRC, that is, the field cages and corncribs. Consequently, the N_e at CS is significantly larger than at the subdivided CNPRC colony. As such, the discrepancy between Fe, which denotes the number of equally contributing founders that would be expected to produce the same level of genetic diversity as the current population, and Fg, which is an estimate analogous to Fe but takes genetic drift into account, highlights that there is more than 3-fold the loss in unique founder alleles due to genetic drift at the CNPRC than in CS, resulting in 4-fold higher GU values in the CS colony (Table 1).⁶⁵

Consistent with these findings, genetic drift, which increases F, is considerably greater at the CNPRC than at CS, and genetic drift and changes in F are proportional to ½N_e.⁵¹ Consequently, at the CNPRC, where animals are housed in indoor facilities, breeding cages, and corncribs, isolated subpopulations have a lower N_e than at CS. Therefore, in addition to increasing the chance of consanguineous matings among captively bred animals, conditions that caused the increase in F also resulted in more pronounced effects of genetic drift in the subdivided CNPRC colony. The estimates of Fe and Fg generated in this study indicated that far fewer founders than expected contributed to the current level of genetic diversity at the CNPRC and CS colonies, ∼50%, compared with the actual number. This supports the earlier observation that reproductive success is limited to a few animals at both colonies due to a positive correlation between male social rank and reproductive success.^24,70

Fg and Fe are the numbers of founders (unrelated animals) that, if bred at random, would produce the same diversity (and thus average MK and GU) as in the current population. While GU, F, and Fe/Fg provide a valuable context on founder representation, drift, and inbreeding, the ANCOVA analyses show that MK uniquely integrates these dynamics into a colony-wide measure of relatedness that can be directly used to preserve genetic diversity. MK is widely regarded as the most practical metric for preserving genetic diversity, since breeding strategies that minimize MK consistently retain more variation than approaches focused only on inbreeding avoidance or founder representation.⁶⁵ Therefore, MK is treated as the key metric for safeguarding the long-term viability of captive colonies in support of biomedical research.

MK is also a measure of an animal’s value to the current living population; it represents the average (mean) relatedness (kinship) of an animal to its population. Although most attention is often given to F, the average MK is more important. Inbreeding is defined in part by kinship, not the other way around. Therefore, MK is a crucial parameter to monitor to maintain genetic diversity within a population. Because MK reflects an individual’s average relatedness to the population, it serves as a predictive tool for identifying animals whose genetic contributions can either dilute or concentrate diversity. Prioritizing low-MK individuals in breeding decisions is therefore essential for sustaining long-term genetic health.

In the present study, there is an increasing trend in MK for both the CNPRC and CS over time. If current breeding strategies remain unchanged at the CNPRC, the increase in MK (and consequently F) is expected to continue. Breeding should preferably involve animals with lower MK to increase population diversity. Ignoring these animals would lead to a continued decline in genetic diversity. For the CS colony, prioritizing the removal of animals with high MK could be part of its population control measures to prevent the loss of genetic diversity. The trends in F, GU, and MK values over a 2-decade period (2000-2024) revealed significant changes in genetic diversity and structure within the colonies. The F values for the CNPRC showed a notable increase over time, characteristic of closed captive-breeding colonies, while the CS values remained relatively stable, as expected in free-ranging and wild populations.²⁴ Both colonies exhibited a general decline in GU values, while MK values showed slight fluctuations, generally increasing as they approached the 2020-2024 period. The CNPRC data revealed notable differences in MK between males and females, suggesting that sex accounted for a small portion of the variation in MK. The CS population also showed a significant difference in MK between males and females. Nonetheless, only a weak negative correlation was observed among estimates of F, MK, and GU at both colonies, and this correlation was not statistically significant.

While probably male-driven, given the high number of families per male in both colonies, Oldt et al. have reported increasing MK levels among several matrilines at the CNPRC colony, consistent with the findings of this study.²⁴ However, there was no significant difference in GU between sexes when controlling for MK.²⁴ In this study, statistically significant differences in GU were observed between males and females at the CNPRC, while no significant relationship was evident between F and GU. The CS data also exhibited comparable patterns to those at the CNPRC, where sex influenced MK and showed a nonsignificant negative relationship between GU and MK.

The CNPRC and CS colonies exhibit distinct genetic and reproductive characteristics, influenced by their respective pedigree histories and management practices. The CNPRC colony exhibited higher levels of inbreeding and experienced much greater effects of genetic drift than did the CS population. The harem system in the corncrib, where the average F across the corncribs (0.05) is 5-fold greater than across field cages (0.01), may have also contributed to the higher inbreeding coefficients observed at the CNPRC (Table 3). These findings underscore the importance of monitoring inbreeding levels to ensure the long-term viability of these populations. Notably, the findings of this study suggest that high levels of inbreeding (as observed in the elevated levels of F and MK) and genetic drift in the CNPRC colony are greater than those of the CS population, despite the multiple historical bottleneck events in the CS population.^32,42,49 Furthermore, the CNPRC has also attempted to promote gene flow within its colony by importing a few thousand animals from various geographic regions over the years for use in the colony’s breeding program.^32,49 For instance, between the early 1980s and late 1990s, the CNPRC breeding colony was supplemented with ∼2,400 Indian-origin rhesus macaques introduced through both direct importation from the Indian subcontinent and secondary breeding centers in the United States and overseas.⁷⁶

During the same time frame, the CNPRC purchased ∼180 animals from China to support comparative biomedical research and enhance genetic variability. To introduce novel alleles into the predominantly Indian-origin colony, Chinese animals were systematically integrated into breeding groups. By 2009, the resulting Chinese–Indian hybrids had grown to constitute one-fifth of the total colony population.^{22,25,32,37,77} While implementing strategies to decrease its Chinese-origin and hybrid animals, the CNPRC bolstered its Indian-origin stock again in July 2021, acquiring 20 CS animals from CS for its breeding program.²³ Despite these efforts, the levels of F and MK at the CNPRC are greater than those of CS. While these efforts introduced novel alleles, the persistence of elevated F and MK values suggests that genetic subdivision and breeding practices continue to constrain gains in diversity.

This pattern of constrained diversity is further underscored by the limited number of founders who actually contributed to the present populations, especially at the CNPRC. While colony records indicate a total of 409 individuals in the founding stock in CS, theoretical estimates of founder representation suggest ∼27%-32% of these initial colony members contributed genetic material to their descendants in the present study cohorts. At the CNPRC, estimates of only 4% to 8% of the founding stock (N = 400) and the subsequently introduced animals (N = 2,580) contributed genetically to the extant population, a vastly smaller number of genetic contributors than those at CS, which contributed to the animals currently alive at the CNPRC.

To understand the implications of these trends for colony management, estimates of population genetic parameters, such as F, GU, and MK, are crucial for developing alternative breeding strategies that aim to minimize genetic diversity loss.^24,49,52 For example, prioritizing individuals with lower MK values for breeding can help conserve genetic diversity. Within breeding cages, MK estimates can be used to equalize reproductive success among males, thereby maximizing the N_e relative to the census size (N) and mitigating risks of genetic bottlenecks, inbreeding, and intra-colony differentiation.^22,25

One such strategy is the infant cross-fostering program, first implemented at the CNPRC in 1983.^59,76 This cost-effective and reliable approach increases the N_e/N ratio by guiding the exchange of infants, particularly females, between mothers housed in different field cages during the first 2 weeks of life.²⁴ This practice promotes colony-wide panmixia, reducing genetic differentiation among breeding units. Notably, female recruits from the program who achieved reproductive success as adults contributed more effectively to matrilineal genetic diversity than do their male counterparts.²⁴

Further refinement of genetic management has involved estimating founder genome contributions using pedigree data, including Fe and Fg. These metrics help evaluate the success of strategies designed to retain and redistribute founder genetic diversity through male-mediated gene flow and infant-swapping protocols.^24,52,65

Pedigree-based genetic management remains a cost-effective method for preserving or enhancing genetic diversity by preventing genetic subdivision and accidental inbreeding. Through parentage assessments, pedigree establishment, individual genetic characterization, and longitudinal monitoring of population genetic parameters, it is possible to maintain the integrity of nonhuman primate research models while meeting the production demands of captive breeding programs.

These findings underscore the need for targeted genetic management strategies that prioritize minimizing inbreeding and preserving founder representation. Future policies should emphasize the routine monitoring of metrics, such as F, GU, and MK, and support interventions, such as strategic mate selection and controlled migration, that enhance genetic diversity and long-term colony viability. As captive colonies continue to support critical biomedical research, integrating pedigree-based metrics into routine management protocols will be essential, not only for preserving genetic integrity but also for ensuring the translational relevance of research outcomes.

Conclusion

Genetic trends at the CNPRC colony raise significant concerns about the long-term sustainability and continued value of this colony to biomedical research. With <10% of founders contributing genetically to the current population, there is clear evidence of reduced genetic diversity, mainly in the form of rare founder alleles. The elevated F values increased genetic drift and diminished gene flow, further highlighting the risks associated with a fragmented population structure. Despite efforts to equalize reproductive success through cross-fostering and introducing new individuals, these strategies have not been sufficiently effective in reversing the loss of founder representation or mitigating inbreeding.

To address these challenges, the CNPRC must adopt a more comprehensive genetic management framework. The CS colony, characterized by its blend of naturalistic social dynamics and managed conditions, offers a compelling model for developing ecologically informed strategies to preserve genetic diversity and promote long-term population health. This includes revising breeding strategies to enhance founder representation, prioritizing underrepresented lineages, and implementing rotational breeding to increase gene flow and reduce drift.⁶⁹

While active monitoring of MK values should inform breeding decisions, supported by centralized pedigree records and genomic validation, when ranking individuals for breeding purposes, the preservation of rare genotypes should take precedence over minimizing background co-ancestry or kinship. This can be achieved by first prioritizing animals with the highest GU, then ranking them by MK within GU.²⁶ This strategy reflects a deliberate emphasis on maintaining genetic novelty, even at the potential cost of increased relatedness within the colony. Addressing disparities in reproductive success and lifespan through standardized care protocols will further promote a balanced and resilient population structure.

Finally, where feasible, next-generation sequencing should be applied to guide pairings and minimize inbreeding.⁷⁸ MiSeq (Illumina, San Diego, CA) short-read sequencing provides high accuracy, enables detection of small variants, accommodates large sample sets, and offers cost-effectiveness per base.⁷⁹ By providing high-throughput, cost-efficient genomic data, next-generation sequencing enables colony managers to identify, validate, and translate macaque genetic variation more effectively. Collectively, these measures offer a pathway to restoring genetic integrity, preserving diversity, and securing the CNPRC colony’s future as a vital resource for biomedical research.

Conflict of Interest

The author has no conflicts of interest to declare.

Funding

Research Center Base Grant Award no. CNPRC-P51 OD011107, awarded to the Regents of the University of California, and no. CPRC-P40 OD012217-37, awarded to the University of Puerto Rico School of Medicine (Principal Investigator: C. Sariol), supported this study.