Genetic Variability and Population Structure of Camelus from Kazakhstan Inferred from 17 STR Markers

Abstract

Camels have been essential to human survival and development across the arid Central Asian steppes, particularly in Kazakhstan, where the breeding of one-humped and two-humped camels is a longstanding tradition supporting the nomadic lifestyle. This study aimed to assess the genetic diversity and population structure of these camels across their distribution range in Kazakhstan. Blood samples from 100 individuals were collected from five locations, Almaty (ALA), Atyrau (ATR), Shymkent (SHK), Kyzylorda (KZL), and Taraz (TRZ), and genotyped using 17 microsatellite markers. All loci were polymorphic, with a mean observed heterozygosity of 0.707 in C. dromedarius and 0.643 in C. bactrianus. The highest expected heterozygosity (He = 0.939) was observed at VOLP67 in C. bactrianus and at VOLP03 in C. dromedarius. Genetic differentiation was low (FST = 0.021), indicating a weak population structure between the two species with substantial gene flow (Nm = 19.972). The hybrid analysis identified 31% hybrids, including F1, F2, and backcrosses, with the highest frequencies in KZL and TRZ, moderate frequencies in ATR, and lowest frequencies in SHK and ALA. These patterns, consistent with STRUCTURE clustering, reflect widespread but regionally variable hybridization. The phylogenetic analysis revealed three clades, separating Bactrian camels (ALA), dromedaries (SHK), and a hybrid group (ATR, KZL, and TRZ). These findings enhance our understanding of the genetic diversity of Kazakhstan’s camels and support effective conservation, breeding strategies, and genotyping applications in camel husbandry.

1. Introduction

Camels (Camelus spp.) are iconic livestock species that hold immense cultural, economic, and livelihood significance for many communities across Africa, the Middle East, and Asia. As highly adapted animals capable of thriving in harsh, arid environments, camels have provided vital transportation, food, and other essential resources to pastoralist societies for centuries. Three distinct camel species are recognized: the one-humped dromedary (Camelus dromedarius), the domesticated two-humped Bactrian camel (Camelus bactrianus), and the wild camel (Camelus ferus) [1]. Dromedaries are mainly distributed across northern Africa, the Middle East, parts of Asia, and the Indian subcontinent. The wild Bactrian camel is restricted to northwest China and Mongolia, particularly within the Great Gobi Strictly Protected Area, while the domesticated Bactrian camel is found in inner, central, and eastern Asia, including Kazakhstan, Kyrgyzstan, Turkmenistan, Afghanistan, northern Iran, India, Pakistan, and eastern Turkey [2,3].

Kazakhstan, situated at the heart of Central Asia, encompasses vast arid and semi-arid land masses with a continental climate, making it a suitable environment for the breeding of drought-resilient animals like camels. The domesticated two-humped Bactrian camels primarily inhabit the southern and western regions of Kazakhstan, while the dromedary populations, represented solely by the Turkmen “Arvana” breed, reside the southern margins of the country, which is considered the northernmost part of the dromedary’s range globally. The Arvana camel is the only dromedary breed that has been hybridized with the Bactrian camel in Kazakhstan [4,5]. As one of the most important livestock units in the country, camels are primarily bred for milk and meat production, which are rich in nutrients and vitamin C [6].

Over the past century, the camel population in Kazakhstan has undergone significant fluctuations. In 1927, when Kazakhstan was part of the former Soviet Union, the camel population was estimated to be around 1.2 million. However, after the “collectivization” governmental reforms in 1940, the number of camels decreased to just 100,000 [7,8]. After gaining independence, Kazakhstan underwent agricultural reforms, which led to the rapid development of traditional camel breeding systems in the country. As a result, the camel population in Kazakhstan has stabilized, reaching 260.5 thousand by 2022 [9].

Numerous genetic studies on camels have been conducted worldwide, including assessments of genetic diversity and population structure to investigate local or regional camel groups, as well as phylogenetic studies examining the demographic history of camel populations [10,11,12,13,14]. Additionally, there are studies that explore the associations between phenotypic traits and underlying genotypes, which can inform marker-assisted breeding programs or adaptations to specific environments [15].

However, the camel populations in Kazakhstan have been poorly investigated, representing a research gap. Previously, only a few studies have been dedicated to the examination of camel breeds in Kazakhstan, employing various approaches [16,17,18]. Elmira et al. [16] investigated the allele pool and genetic diversity of dairy camel populations in South Kazakhstan, specifically the Arvana and Kazakh Bactrian camel breeds, using 12 STR microsatellites. Later, the genetic structure of Bactrian camels was examined based on the D-loop region of mitochondrial DNA [17], as well as the polymorphism of casein genes in Kazakh camel breeds [18]. More recently, Amandykova et al. [19] analyzed genome-wide SNP data from five F1 hybrids on a single farm in the Almaty region, comparing them with publicly available parental genomes. While their study provided useful insights into hybridization at the genomic level, its limited sample size and geographic coverage did not allow for a broader evaluation of the genetic structure, regional variation, or admixture beyond first-generation hybrids.

To complement previous work, this study analyzes a broader and geographically representative dataset of 100 camels sampled from five major camel-breeding regions in Kazakhstan. We employed 17 polymorphic microsatellite markers (short tandem repeats, STRs), which are commonly used in livestock genetics due to their high informativeness, cross-species utility, and suitability for assessing genetic diversity, population structure, admixture, and inbreeding. In camelids, STR-based studies have successfully revealed interspecific variation between C. dromedarius and C. bactrianus, as well as among South American camelids such as alpacas and llamas, along with patterns of gene flow across regions. Given their practicality and cost-efficiency, STRs remain a useful tool in camel breeding and conservation programs. This work provides the first comprehensive STR-based genetic reference for Kazakh camel populations and contributes novel insights into species admixture and gene flow under traditional husbandry systems.

In this study, we explored the genetic diversity and population structure of Kazakhstani one- and two-humped camels from five regions of Kazakhstan using a panel of 17 microsatellite markers originally developed for various camelid species. These markers have demonstrated effective cross-species amplification and are widely used in genetic studies within the Camelus genus. We hypothesized that the spatial distribution and traditional breeding practices may have influenced the genetic structuring of these camel populations. Specifically, this study aimed to (a) assess the genetic diversity and population structure of C. dromedarius and C. bactrianus separately, and (b) conduct a pooled analysis of regional populations to evaluate the extent of genetic admixture and assess the effectiveness of the STR marker panel for species delimitation.

2. Materials and Methods

2.1. Sampling Information

A total of 100 samples of peripheral blood were taken from one-humped Camelus dromedarius and two-humped camel Camelus bactrianus as the research material. They represent five populations of camels throughout Kazakhstan (

Table 1. Sampling information of Kazakhstani Camelus used in this study.

Location, Farm	(Sample Size), Taxon
(ATR) Atyrau, WKZ	(n = 20), C. dromedarius
(KZL) Kyzylorda, SWKZ	(n = 10), C. dromedarius; (n = 10), C. bactrianus
(SHK) Shymkent, SKZ	(n = 20), C. dromedarius
(TRZ) Taraz, SKZ	(n = 10), C. dromedarius; (n = 10), C. bactrianus
(ALA) Almaty, SEKZ	(n = 20), C. bactrianus

W—west; S—south, SW—southwest; SE—southeast; KZ—Kazakhstan.

; Figure 1): the Atyrau region (ATR)—C. dromedarius (n = 20); Kyzylorda region (KZL)—C. dromedarius (n = 10) and C. bactrianus (n = 10); Turkestan region, Shymkent city vicinity (SHK)—C. dromedarius (n = 20); Zhambyl region, Taraz city vicinity (TRZ)—C. dromedarius (n = 10) and C. bactrianus (n = 10); and Almaty region (ALA)—C. bactrianus (n = 20).

All individuals were adults, visually identified as one or two-humped camels, and hybrids were not included to maintain homogeneity within the samples. The collection of blood samples was carried out under the supervision of a veterinary specialist in accordance with the program of injections from camels, without causing any stressful situations. Blood samples were collected by sterile needles in vacuum test tubes containing the EDTA reagent. The collected biomaterials were delivered to the laboratory using special transportable freezer boxes at a temperature of ±4 °C.

2.2. Laboratory Protocol

A total of 17 microsatellite markers were selected and used for typing, among those recommended by the International Society of Animal Genetics (ISAG) and the FAO working group [20] (Table S1). These loci—CMS13, CMS15, CMS50, LCA63, LCA66, YWLL08, YWLL09, YWLL38, YWLL44, CVLR05, CVLR06, CVLR07, VOLP03, VOLP08, VOLP10, VOLP32, and VOLP67—were originally developed for various camelid species, including Camelus dromedarius, Camelus bactrianus, and South American camelids (e.g., Lama glama and Vicugna pacos), and have shown reliable cross-species amplification [21,22,23,24,25]. Microsatellites were divided into four multiplexes (M) according to the dye used, annealing temperature, and allelic range (M1—VOLP03, VOLP08, LCA66, and CVRL05; M2—YWLL09, YWLL38, YWLL44, CMS15, CVLR07, and VOLP32; M3—CMS50, LCA63, VOLP67, and YWLL08; and M4—VOLP10, CMS13, and CVRL06).

Genomic DNA was extracted from the collected biomaterials using the commercial GeneJET Genomic DNA Purification Kit (Thermo Scientific, Waltham, MA, USA) according to the manufacturer’s standard protocol. The polymerase chain reaction (PCR) was prepared in a specialized UV-sterilized box to ensure contamination-free conditions. Commercially available PCR reagent kits were utilized for the analysis (PCR Master Mix, Thermo Fisher Scientific, USA). The total reaction volume was 25 µL, comprising 0.5–1.5 µL of genomic DNA at a concentration of 50–60 ng/µL, 1 µL of each forward and reverse primer (from 10 µM stocks, resulting in a final concentration of 0.4 µM per primer), 12.5 µL of PCR Master Mix (containing 0.05 U/µL Taq DNA polymerase, reaction buffer, 4 mM MgCl₂, and 0.4 mM dNTPs, including dATP, dCTP, dGTP, and dTTP), and nuclease-free water to complete the volume.

The PCR amplification protocol contained an initial denaturation at 95 °C for 3 min, followed by 40 cycles of denaturation at 95 °C for 30 s, annealing for 30 s at 55 °C or 60 °C depending on the primer, elongation at 72 °C for 1 min, and a final extension at 72 °C for 12 min with subsequent cooling to 4 °C. The PCR reaction was performed using the GenAmp 2700 (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA) thermal cycler. The PCR fragments were resolved on an ABI Prism 3500 (Applied Biosystems, Waltham, MA, USA) fragment analyzer, and the data were analyzed using the GENEMAPPER 4.1 software (Applied Biosystems). In order to avoid any uncertainty about the size of amplified fragments and to correctly identify the different alleles for each individual, three operators validated the fluorescence peak obtained with GENEMAPPER 4.1 software.

2.3. Genetic Structure Analysis

Polymorphic information content (PIC) [26] was calculated by POWERMARKER (version 3.25) [27] for each microsatellite marker, species, and population. The number of alleles, expected (He) and observed (Ho) heterozygosity [28], Shannon’s information index (I), Wright’s fixation indices—FIS (inbreeding coefficient within subpopulations, indicating the reduction of heterozygosity due to non-random mating), FST (fixation index, measuring genetic differentiation among subpopulations due to genetic drift), and FIT (overall inbreeding coefficient relative to the total population)—were calculated according to the method of Weir and Cockerham (1984) [29] using FSTAT software (version 2.9.3.2) [30]. The effective number of migrants (Nm) as well as the pairwise population matrix of Nei’s standard genetic distance (DA) were calculated using GenAlex 6.5 and Excel microsatellite toolkit (version 3.1) software [31]. Deviation from Hardy–Weinberg equilibrium (HWE), as well as the private allele number, were quantified in GENEPOP v4.7.5 [32,33]. Phylogenetic relationships were assessed using Nei’s genetic distances following the neighbor-joining method of Saitou and Nei (1987) [34] implemented in MEGA version 7.0 [35]. To evaluate the bootstrap support of the clades, we applied the method of Felsenstein (1985) [36], conducting 1000 bootstrap replications to assess the stability and reliability of the inferred tree topology. A factorial correspondence analysis (FCA) was conducted using GENETIX version 4.03 [37] and each marker specific allele frequency to identify genetic correlations between groups.

To disentangle intra-species population structure from inter-species admixture, we analyzed our microsatellite data using three complementary datasets. The regional (mixed-species) dataset included all 100 individuals from the five sampling sites (see Table 1). This dataset, which contained both species at locations where they naturally co-occur (Kyzylorda (KZL) and Taraz (TRZ)), was used to visualize overall genetic clustering and to detect admixed individuals.

In addition, we analyzed two species-specific datasets. The dromedary-only dataset included 60 phenotypically unambiguous one-humped camels from Atyrau (ATR), Shymkent (SHK), KZL-D, and TRZ-D. The Bactrian camel-only dataset comprised 40 two-humped camels from Almaty (ALA), KZL-B, and TRZ-B. In both cases, individuals of the opposite species were excluded in order to avoid ΔK inflation and to preserve the resolution of population structure within each species.

All datasets were analyzed using STRUCTURE v2.3.4 [38,39]. For each analysis, twenty independent runs were performed at each K value (ranging from K = 2 to 10), with 1,000,000 Markov chain Monte Carlo (MCMC) iterations following a 250,000-step burn-in. STRUCTURE HARVESTER v0.6.94 [40], implemented in Python v3 [41], was used to calculate the ad hoc ΔK statistic according to the method of Evanno et al. [39] to determine the most likely number of genetic clusters. CLUMPP v1.1.2 [42] was used to align the replicate runs, and the final bar plots were visualized using DISTRUCT v1.1 [43].

To estimate the proportions of hybridization among populations of both species of camels and identify the genetic composition of individuals based on multilocus genotype data, we performed a Bayesian framework analysis using NewHybrids software (version 1.1) [44]. Missing data were denoted using the placeholder (−9). The analysis was conducted using unlinked loci, ensuring that loci met Hardy–Weinberg equilibrium and linkage disequilibrium assumptions within parental populations. Default priors for allele frequencies (“Jeffreys”) and mixing proportions were used. The Markov Chain Monte Carlo (MCMC) simulations were run for 100,000 iterations following a burn-in period of 20,000 iterations. Convergence was assessed by examining posterior probabilities and parameter stability across multiple runs. Individuals were assigned to the genetic classes with posterior probabilities ≥0.9. This program implements a Bayesian framework to assign individuals to distinct genetic classes based on their multilocus genotype. The classes include the pure parental populations P1 and P2, F1 hybrids, F2 hybrids, and backcrosses to each parent (P1 × F1) and (P2 × F1), respectively.

3. Results

3.1. Genetic Polymorphisms and Diversity

3.1.1. Species-Level Genetic Diversity

A total of 244 alleles were identified in 60 C. dromedarius samples and 241 alleles in 40 C. bactrianus samples across 17 microsatellite loci, averaging 14.35 ± 1.66 and 14.18 ± 1.56 alleles per locus, respectively (Table S2). When pooled by region (n = 100), the five regional populations revealed 150 alleles, with a mean of 8.80 ± 0.42 alleles per locus (

Table 2. Characteristics of 17 microsatellites across pooled five regional camel populations (n = 100) from Kazakhstan.

Locus	Na	Ne	I	Ho	He	uHe	FIS	FIT	Fst	Nm	PIC	HWE
CVRL05	8.400	5.035	1.787	0.747	0.791	0.811	0.055	0.120	0.068	3.418	0.824	ns
LCA66	10.600	5.781	2.025	0.677	0.826	0.848	0.181	0.262	0.099	2.272	0.917	*
VOLP08	7.800	3.867	1.561	0.614	0.721	0.741	0.148	0.275	0.149	1.425	0.824	ns
VOLP03	13.000	9.063	2.304	0.603	0.871	0.900	0.308	0.359	0.073	3.168	0.941	***
CMS15	9.000	5.179	1.798	0.649	0.775	0.798	0.163	0.256	0.111	1.999	0.855	**
YWLL09	4.200	2.622	1.112	0.472	0.601	0.617	0.214	0.298	0.106	2.106	0.590	ns
VOLP32	4.400	3.363	1.265	0.650	0.686	0.704	0.053	0.114	0.065	3.593	0.736	ns
YWLL38	6.600	3.526	1.490	0.710	0.713	0.731	0.004	0.059	0.055	4.256	0.716	ns
CVRL07	6.400	3.442	1.417	0.430	0.690	0.708	0.377	0.456	0.127	1.722	0.787	***
YWLL44	7.200	4.165	1.599	0.759	0.745	0.764	−0.019	0.048	0.066	3.520	0.763	ns
CMS50	12.800	8.121	2.248	0.800	0.860	0.882	0.070	0.136	0.071	3.288	0.923	ns
LCA63	5.200	3.759	1.424	0.770	0.724	0.742	−0.064	0.013	0.073	3.194	0.760	ns
VOLP67	15.200	10.132	2.468	0.809	0.895	0.918	0.095	0.141	0.051	4.697	0.940	***
YWLL08	12.600	6.905	2.162	0.860	0.846	0.867	−0.017	0.082	0.097	2.322	0.933	ns
VOLP10	12.600	6.005	2.112	0.808	0.822	0.844	0.017	0.105	0.090	2.536	0.903	*
CMS13	8.600	5.158	1.843	0.790	0.802	0.822	0.015	0.079	0.065	3.570	0.840	ns
CVRL06	5.000	3.402	1.315	0.450	0.658	0.675	0.316	0.458	0.207	0.957	0.804	***
Mean	8.800	5.266	1.761	0.682	0.766	0.787	0.113	0.192	0.093	2.826	0.827	ns

Number of observed alleles (Na), number of effective alleles (Ne), Shannon’s information index (I), observed heterozygosity (Ho), expected heterozygosity (He), unbiased expected heterozygosity (uHe), inbreeding coefficients (F_IS), (F_IT), and fixation index (F_ST) p ≤ 0.001, (PIC) polymorphic information content, (Nm) number of migrants, Hardy–Weinberg equilibrium (HWE) statistics. Significance levels for HWE: ns—not significant, * p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001.

). All loci were polymorphic (p ≥ 0.05), indicating substantial genetic diversity. Mean statistical indices for both types of data, species-specific and population-specific, based on 17 STR loci are summarized in Table S3.

Loci VOLP67 and VOLP03 were particularly informative in C. dromedarius (Na = 28; Ne = 16.31), while VOLP67 and YWLL08 showed high polymorphism in C. bactrianus (Na = 25; Ne = 16.50 and 14.68). The mean observed heterozygosity (Ho) was similar in both species: 0.707 in dromedaries and 0.643 in Bactrian camels (Table S3). The highest He was 0.939 in both species, observed at VOLP67 (C. bactrianus) and VOLP03 (C. dromedarius). The inbreeding coefficient (FIS) ranged from –0.121 to 0.363 in dromedaries and –0.180 to 0.517 in Bactrian camels (Table S2), while the mean values composed 0.143 and 0.226, respectively (Table S3). Hardy–Weinberg equilibrium (HWE) tests showed significant deviations across all 17 loci in C. dromedarius and at 12 loci in C. bactrianus, suggesting possible population substructure or selection [45,46]. The private allele analysis revealed that both species harbored low-frequency alleles unique to their gene pools. The mean frequencies of private alleles ranged from 7 to 10%, with the highest frequency (~34.6%) observed at VOLP03 in C. bactrianus.

3.1.2. Regional Population-Level Genetic Diversity

An analysis of five regional populations (ALA, ATR, KZL, SHK, and TRZ; n = 100 pooled samples) yielded 150 alleles (mean Na = 8.80 ± 0.42 per locus). The Shannon diversity index was highest in ALA (I = 1.879 ± 0.134) and lowest in SHK (I = 1.633 ± 0.096). VOLP67 (I = 2.468), followed by VOLP03, were the most informative markers across all populations. Observed heterozygosity (Ho) was highest in SHK (0.749 ± 0.047) and ATR (0.719 ± 0.033), and lowest in ALA (0.628 ± 0.065), despite its higher allelic richness. This indicates possible substructuring or inbreeding in ALA. Inbreeding coefficients were highest in ALA (F = 0.204), followed by KZL and TRZ, with minimal values in SHK (F = 0.001) and ATR (F = 0.050). Private allele frequencies varied, with ALA standing out (range: 0.013–0.788), while other populations had narrower ranges (~0.025–0.225), supporting both a shared ancestry and region-specific divergence.

3.2. Genetic Structure and Differentiation

To evaluate genetic structure and differentiation within and between camel species and regional populations in Kazakhstan, we analyzed F-statistics, AMOVA, genetic distances, and clustering patterns.

In C. bactrianus, the highest levels of differentiation were observed at loci CVRL06 (FST = 0.239) and VOLP08 (F_ST = 0.230), with the greatest inbreeding coefficients (F_IS = 0.517; F_IT = 0.611) detected at CVRL07. Several loci, such as YWLL44, VOLP08, and YWLL38, showed negative F_IS values, indicating excess heterozygosity. In C. dromedarius, locus CMS15 exhibited the highest F_ST (0.160), while VOLP03 showed the highest inbreeding levels (F_IS = 0.363; F_IT = 0.409). Across all populations, CVRL06 had the highest F_ST (0.207), while other loci, including YWLL44 and LCA63, presented heterozygote excess.

The analysis of molecular variance (AMOVA) showed that 77% of the total genetic variation was within individuals, 15–20% among individuals, and only 8% among regional populations (F_ST = 0.083; p ≥ 0.001). When grouped by species, just 3% of the variation was attributable to differences between C. dromedarius and C. bactrianus (F_ST = 0.028; p ≥ 0.001), indicating minimal interspecies differentiation (

Table 3. Analysis of molecular variance of Kazakhstan camel populations.

Source of Variation	d.f.	SS	MS	Est. Var.	% of Var.
Five regional populations of camels
Among population	4	128.580	32.145	0.608	8%
Among individuals	95	743.075	7.822	1.076	15%
Within individuals	100	567.000	5.670	5.670	77%
Total	199	1438.655		7.354	100%
FST = 0.083 (p ≥ 0.001); FIS = 0.159 (p ≥ 0.001); FIT = 0.229 (p ≥ 0.001)
Two species of camels
Among species	1	27.777	27.777	0.203	3%
Among individuals	98	844.338	8.616	1.475	20%
Within individuals	100	566.500	5.665	5.665	77%
Total	199	1438.615		7.344	100%
FST = 0.028 (p ≥ 0.001); FIS = 0.207 (p ≥ 0.001); FIT = 0.229 (p ≥ 0.001)

d.f.—degrees of freedom; SS—sum of squares; MS—mean of squares; Est. Var.—estimated variance.

). This low divergence, supported by similar allele frequencies and low per-locus F_ST values, suggests historical or ongoing gene flow, likely due to widespread hybridization practices in Kazakhstan.

Nei’s genetic distance (DA) between the two camel species revealed a moderate divergence (0.028), while Wright’s FST index was 0.021, indicating low genetic differentiation—likely due to ongoing gene flow or shared ancestry. Among regional populations, FST values ranged from 0.041 to 0.088, suggesting a more pronounced population structure or restricted gene flow. The corresponding DA values ranged from 0.339 (ATR–KZL) to 0.854 (ALA–ATR), reflecting considerable variation in genetic distance between populations (

Table 4. Pairwise comparisons of genetic differentiation and distance among Kazakhstani camel populations. Wright’s fixation index (FST) is shown above the diagonal, and Nei’s genetic distance (DA) is shown below the diagonal.

Pop	ALA	ATR	SHK	KZL	TRZ
ALA		0.088	0.075	0.064	0.065
ATR	0.854		0.059	0.041	0.050
SHK	0.615	0.466		0.052	0.062
KZL	0.555	0.339	0.429		0.044
TRZ	0.574	0.433	0.553	0.385

).

An unrooted neighbor-joining tree grouped ATR and KZL into a single clade, while two-humped camels represented by the population (ALA) forming a distinct branch (Figure 2). The factorial correspondence analysis (FCA) supported this structure: the species-specific FCA separated individuals into two clusters defined entirely by Axis 1, while the regional FCA revealed three clusters, with SHK and ALA appearing genetically distinct, and ATR, KZL, and TRZ forming a single admixed group (Figure 3).

Bayesian clustering using STRUCTURE indicated an optimal K of 5 for C. dromedarius (ΔK = 12.05) and K = 3–4 for C. bactrianus (Figure 4), and with regional populations showing the strongest support at K = 5 and highest likelihood at K = 10 (Figure S1).

The NewHybrids analysis identified various hybrid classes across the sampled populations. The highest hybrid frequencies were observed in Kyzylorda (KZO) and Taraz (TAR), with moderate levels in Atyrau (AT), and lower frequencies in Shymkent (SH) and Almaty (ALA) (Figure 5). Using a ≥90% posterior probability (PP) threshold for confident classification, from 100 sampled camels, 31 were identified as hybrids, comprising eight samples of each hybrid type (F1 and F2), respectively. Bactrian backcrosses (P1 × F1 = 13 samples) and dromedary backcrosses (P2 × F1 = 2 samples) were observed. Pure parental types were represented by 13% Bactrian camels (P1) and 19% dromedary camels (P2), while 37% of individuals remained unclassified with the PP < 90 (Table S4).

The STRUCTURE analysis corroborated these findings by revealing mixed ancestry components in the same populations where NewHybrids detected hybrids. The agreement between both methods confirms extensive intra-population genetic diversity, minor regional differentiation, and a notable prevalence of hybridization—particularly in regions where both C. dromedarius and C. bactrianus are managed together.

4. Discussion

This study provides the first comprehensive microsatellite-based assessment of genetic diversity and population structure in Kazakhstani C. dromedarius and C. bactrianus. Using a panel of 17 polymorphic STR markers, we aimed to investigate whether the spatial distribution and traditional breeding practices, including historical hybridization, have shaped genetic differentiation across five regional populations. The results confirm high levels of allelic diversity, a moderate population structure, and evidence of interspecific gene flow, reflecting the complex demographic and breeding history of camels in central Asia.

Consistent with the expectations from previous studies, both dromedary and Bactrian camels exhibited substantial allelic richness, with over 240 total alleles detected in each species [47,48,49,50]. The average number of alleles per locus was comparable between species (14.35 in C. dromedarius and 14.18 in C. bactrianus) and exceeded the values reported for camel populations in parts of North Africa [51], India [52], and Sudan [53]. Certain loci, such as VOLP67 and VOLP03, proved especially informative, showing high allelic diversity and PIC values (>0.93) comparable to or exceeding those found in Saudi Arabian [13] and Australian camels [54]. These findings indicate that Kazakh camels, despite their localized distribution, retain broad genetic variability that may reflect both historical selection and limited artificial breeding pressure.

Genetic diversity within populations was further supported by high heterozygosity values (Ho = 0.707 in dromedaries and 0.643 in Bactrian camels), as well as strong polymorphism across all loci. The observed and expected heterozygosity values were consistent with or exceeded those reported in African [55] and Middle Eastern camel populations. The Shannon information index (I) was slightly higher in C. bactrianus, indicating marginally greater allelic evenness in this species. These results confirm that both species maintain genetically diverse populations within Kazakhstan. Among the studied loci, YWLL09 exhibited the lowest He value, while the highest values were observed at VOLP67 and VOLP03 (He = 0.939). The average He across all camel populations (0.766) was slightly lower than that reported in Iranian camels (He = 0.861) [56]. Additionally, the polymorphic information content (PIC) values in our study exceeded those reported for Pakistani Mareecha and Barela breeds (0.72 and 0.70, respectively) [57].

Despite this diversity, our findings revealed moderate levels of genetic differentiation between regional populations (FST = 0.083) and minimal species-level divergence (FST = 0.028). These values align with previous reports of a weak interspecies structure due to hybridization [47], but are higher than those observed in panmictic or extensively interbred camel populations elsewhere. Notably, AMOVA indicated that 77% of the total variation resides within individuals, with only 8% attributable to population-level differences. This underscores the extensive shared genetic background among camels from different regions. The minimal interspecies divergence supports long-term gene flow between dromedaries and Bactrian camels, likely facilitated by Kazakhstan’s tradition of hybrid breeding—intended to combine endurance and strength. This is consistent with the identification of F1, F2, and backcross hybrids in the NewHybrids analysis.

Although our study focused on neutral genetic variation, the hybrid classes we detected have clear relevance to production. Traditional Kazakh breeders have long exploited the nar (F1) and subsequent backcross generations to capture heterosis for milk yield, body size, and endurance under continental conditions. Field reports indicate that F1 animals can surpass both parental species in daily milk production and growth rate, while backcrosses are favored for combining the dromedary’s heat tolerance with the Bactrian camel’s cold resilience (e.g., Faye and Bengoumi 2018; Raziq 2008) [58,59]. The geographic pattern we observed—higher frequencies of F1 and F2 hybrids in Kyzylorda and Taraz, and lower frequencies in Almaty and Shymkent—matches the distribution of farms that specifically select for higher-yield dairy hybrids. Although we did not collect production data in this study, linking milk yield and growth traits with the genetic groups identified by STRUCTURE and NewHybrids would be a useful next step. Future genome-wide association studies (GWAS) on hybrids could help identify genes linked to productivity and support better-informed breeding programs in Kazakhstan.

Multivariate analyses further supported the population-level substructure. FCA revealed three regional clusters, distinguishing ALA (Bactrian camels), SHK (dromedaries), and a mixed cluster comprising ATR, KZL, and TRZ populations. These patterns correspond to the geography and known hybrid zones. The STRUCTURE analysis suggested a subpopulation structure at K = 5 in dromedaries and regional populations, and K = 3 in Bactrian camels, consistent with moderate spatial partitioning and localized breeding systems. Neighbor-joining phylogeny based on Nei’s genetic distances positioned ALA and SHK as genetically distinct, while central populations formed an admixed group, reflecting the influence of pastoralism, regional mating, and moderate inter-population connectivity.

Fixation indices also pointed to within-population structure. FIS values ranged from slight inbreeding to excess heterozygosity at certain loci, suggesting balancing selection or localized outcrossing. When comparing species, C. bactrianus showed a higher mean FIS (0.126) than C. dromedarius (0.080), indicating a greater tendency toward inbreeding or non-random mating within its populations. This difference may reflect distinct breeding practices: C. bactrianus is often bred in smaller, more isolated herds in mountainous or continental regions, potentially leading to higher inbreeding levels [60,61]. In contrast, C. dromedarius, which is more commonly managed in larger, mobile herds across wider desert areas, may experience more gene flow and outcrossing, thus maintaining lower FIS values [62,63]. These patterns underscore how traditional management systems can shape the genetic structure of camel populations. Loci such as YWLL44 and LCA63 showed consistent deviations from Hardy–Weinberg equilibrium, which could reflect technical or evolutionary causes such as assortative mating or a cryptic substructure [64,65].

Notably, ATR and KZL populations showed the lowest genetic distance (DA = 3.3%), likely due to geographical proximity across arid plains, enabling gene flow during seasonal migrations. Gene flow estimates (Nm = 2.8) confirmed moderate migration, supporting the preservation of shared alleles across populations. This is lower than estimates reported in Saudi Arabia (Nm = 39 in Sofr/Shual, 14.5 in Magaheem/Maghateer) [66], and contrasts with low differentiation in South African [47], Saudi [67], and Sudanese [53] camels. However, our data are congruent with findings of Xiaohong et al. [68], who identified a substructure among Bactrian camel populations in China and Mongolia shaped by natural barriers.

Finally, our findings align with an earlier ISSR-PCR analysis in Kazakhstan, which also revealed the genetic distinction of the SHK dromedary population and admixture in other groups [69]. These complementary results reinforce the need for expanded genomic studies, such as whole-genome re-sequencing, to further explore evolutionary history, adaptive loci, and trait-associated variation. Such insights will support evidence-based breeding programs and help conserve Kazakhstan’s unique camelid heritage.

5. Conclusions

This study used 17 microsatellite markers to assess the genetic diversity and structure of one- and two-humped camel populations across five regions of Kazakhstan. Clear genetic differentiation was observed, particularly between the Almaty C. bactrianus and Atyrau C. dromedarius populations. Evidence of admixture among the Taraz, Kyzylorda, and Atyrau populations suggests ongoing gene flow, likely driven by traditional hybrid breeding practices and geographic proximity. The STR panel effectively revealed both interspecific and regional population structures, and the detection of F1, F2, and backcross hybrids confirms the genetic footprint of traditional hybridization systems in Kazakhstan. These results highlight the need for conservation and breeding strategies to preserve genetic diversity and long-term population viability. Future studies using genome-wide data are recommended to further explore selection, adaptation, and demographic history.