Genome-wide analysis reveals differential admixture dynamics and historical demographic contractions in African cattle

Abstract

The rich cattle genetic diversity in Africa has been supporting livelihoods for centuries, providing food and nutrition security, income, and socio-cultural values. Understanding the genetic architecture of African cattle is essential for conservation and sustainable breed improvement. This study assessed genome-wide diversity, population structure, admixture patterns, and gene flow in 44 cattle populations across West, East, and Southern Africa, utilizing 63,655 medium-density SNP markers. Comparative analyses incorporated South Asian zebu and European taurine breeds for reference. Genetic diversity, estimated via observed heterozygosity and inbreeding coefficients (F_ROH), indicated higher variability in African zebu and admixed populations compared to African taurine cattle. Trypanotolerant taurine breeds such as N’Dama and Lagunaire exhibited low observed heterozygosity and high inbreeding, consistent with historical isolation and natural selection in tsetse-endemic regions. Principal component analysis and ADMIXTURE clustering revealed region-specific patterns of taurine-indicine admixture. West African taurine breeds retained high taurine ancestry, while West and East African zebu breeds showed extensive indicine introgression. Southern African cattle exhibited moderate admixture, with Nguni cattle retaining a distinct Sanga genetic signature. Effective population size (Ne) estimates revealed historical bottlenecks across all regions, temporally coinciding with the nineteenth century rinderpest pandemic. Contemporary Ne values indicated small effective sizes in several taurine and localized zebu populations, warranting conservation attention. This study provides genomic insights into patterns of admixture, gene flow, and demographic variations shaping African cattle diversity. The findings emphasize the need for regionally tailored breeding and conservation strategies to sustain the genetic diversity and adaptive potential of African cattle under evolving climatic and production challenges.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36562-7.

Introduction

Cattle breeds in Africa comprise more than 150 local types, ranging from pure taurine to nearly pure indicine, adapted to various climatic conditions¹. These breeds are uniquely adapted to their environments, with traits such as resilience to hot and arid environment, resistance to parasitic diseases like trypanosomiasis, drought resistance, and the ability to thrive on sparse and low-quality vegetation ^2–4. Their productivity is generally low compared to commercial beef and milk breeds. To improve productivity, smallholder farmers often practice crossbreeding and genetic introgressions using exotic breeds, predominantly the humped zebu cattle⁵. Thus, it is particularly important to develop baseline genetic information and characterize current breeds and their production environments. Characterization of cattle involves the documentation of phenotypic and genetic characteristics⁶, the studies of which are limited in various African cattle breeds².

The evolutionary history of African cattle provides a rich context for their current diversity. Bos primigenius primigenius and Bos primigenius namadicus, the ancestors of Bos taurus taurus (humpless taurine) and Bos taurus indicus (humped indicine or zebu) cattle respectively, were believed to have inhabited Western Eurasia and South Asia⁷. Apart from these Eurasian and South Asian aurochs, fossil records also indicate the existence of Bos primigenius africanus in Northern Africa that could have contributed to the gene pool of the modern-day African cattle^8,9. Archaeological evidence indicates that the earliest cattle introduced into Africa were humpless taurine, which arrived in two waves: longhorns about 5,000 years BC followed by shorthorns 2,500 years later. These taurine cattle migrated from the northeastern part of Africa to the West and East African regions¹⁰. Subsequent to the spread of taurine cattle across the continent, indicine cattle were introduced, likely in two waves. The first wave of indicine cattle entered through the Horn of Africa, coinciding with the development of Swahili-Arab civilizations along the East African coast in the seventh century AD. The second wave occurred during the nineteenth century, associated with rinderpest epidemics that devastated most of the African cattle population¹¹. Studies suggest that cattle reaching Southern Africa were zebu-introgressed taurine cattle of Sanga type¹¹.

African cattle are phenotypically and genetically diverse with most of them being of indicine type characterized by hump, large size and adaptation to arid, tropical environmental conditions^2,12. A smaller number of breeds, predominantly found in West and Central Africa, are of indigenous African taurine type and are humpless and small in size. African taurine cattle are better adapted to the hot, vector-borne disease-prone conditions of West Africa but have lower productivity (e.g., growth rate, suitability to transhumance system) as compared to indicine. Crossbreeding between these groups has resulted in composite breeds like Sanga (taurine-zebu crosses) and Zenga (Sanga-zebu crosses), that have significant zebu genetic background¹². The spatial distribution of these groups reflects Africa’s ecological diversity. Zebu cattle dominate the Western and Eastern parts of Africa, Sanga are predominant in Eastern and Southern Africa, while Zenga are largely confined to East Africa¹³. The highest population numbers of cattle are found in Eastern Africa, compared to the West and Southern regions. Despite the diversity, all African cattle carry taurine mitochondrial DNA (mtDNA), indicating the absence of pure indicine maternal lines in the region². The origins of African cattle remain debated; some studies suggesting domestication within Africa citing the presence of unique mtDNA haplogroups¹⁴, while most others point to migration and colonization subsequent to domestication in the Fertile Crescent⁷. Recent studies using genome-wide single nucleotide polymorphisms (SNPs) indicate a genetic background originating from the African aurochs through male-mediated introgression¹⁵.

The unique adaptive traits found in African cattle include their ability to cope with transhumance pastoralism, inadequate feed supply, high temperatures, and internal and external parasites^15,16. These traits, along with morphological, physical, and biological features of African taurine and indicine cattle, are a result of both natural and artificial selection. Resilient breeds such as the Angoni in Zambia, which adapts to browsing during the dry season, and the Nguni in South Africa, known for its high fertility and early sexual maturity, are examples of these adaptations. Western and Central African taurine breeds such as N’Dama, Lagunaire and Lobi exhibit trypanotolerance, thriving in tsetse-infested areas, a major hotspot for African trypanosomiasis. These adaptive traits have not only enabled survival but have also been key drivers of the economic, social, and cultural development of the continent¹⁷. Genomic investigations of African cattle have expanded in recent years, using genome-wide SNP arrays and whole-genome sequence data to describe patterns of genetic diversity, admixture, and demographic history across the continent ^18–23. Most of these studies focused on East African populations, particularly Ethiopian and Ugandan cattle along with selected West African breeds, providing important insights into taurine and indicine ancestries that shaped African cattle. Detailed genomic characterizations have also been undertaken for individual breeds such as Goudali ²⁴ and Nguni ²⁵, highlighting within-breed diversity and local adaptation. Analyses incorporating high-density SNPs and local ancestry inferences have also shown heterogeneity in admixture levels, genetic divergence, and effective population size estimates among East and West African populations, reflecting historical migrations, ecological adaptations, and prevailing management practices. Despite these advances, the genomic landscape of African cattle remains unevenly represented, with major geographical and breed-level gaps, particularly across Southern and large parts of West Africa where many indigenous populations have yet to be characterized. The present study analyses 60 K SNP genotypes from 36 breeds spanning East, West, and Southern Africa, including several previously unreported populations such as Lobi and Gourunssi from Burkina; Pabli and Yakana from Benin; Zebu Peuhl from Burkina and Benin; Barotse, Tonga, Angoni and Baila from Zambia; Arabe Pur and Borroro from Niger; and Arishy, Butana and Kenana from Sudan, etc., thereby generating a more comprehensive and representative genomic overview of African cattle diversity. In the present study, we aimed to (i) evaluate the diversity of African taurine and indicine breeds in the Western, Eastern and Southern Africa region using genome-wide SNP data (ii) estimate inbreeding levels, historic trends and contemporary effective population size (iii) assess population structure and establish current levels of genetic admixture (iv) understand gene flow and migrations between South Asian zebu and African cattle.

Materials and methods

Animal ethics statement

No animals were specifically sampled for the present study. All genotypic data analyzed were derived from existing DNA repository of unrelated cattle available at Animal Production and Health Laboratory, Joint FAO/IAEA Centre, International Atomic Energy Agency, Seibersdorf, Vienna, Austria. As no live animals were handled and no specific biological samples were collected, ethical approval was not required for this study.

Genotyping, and quality control

A total of 2024 cattle DNA samples were analyzed, representing nine West African taurine, ten West African zebu, three West African crossbred, five East African, six Southern African zebu, and two Southern African crossbred populations, along with nine reference populations (five South Asian zebu and four European taurine). The information on breed names and their distribution is described in Table 1. The characteristic features such as coat color, horn type (longhorn or shorthorn), and presence of hump (humped or humpless) of cattle breeds used in the study are presented in Supplementary information S1 file.

Table 1.

Heterozygosity statistics for African taurine and zebu populations.

Region	Country	Name	Code	Samples	HO	HE
West African Taurine (WAT)	Benin	Bourgou	EBG	48	0.342	0.337
	Benin	Lagunaire	ELN	29	0.261	0.271
	Benin	Pabli	EPB	35	0.345	0.325
	Benin	Somba	ESO	45	0.318	0.309
	Burkina Faso	Gourounssi Nahouri	FGN	27	0.316	0.313
	Burkina Faso	Gourounssi Tenado	FGT	46	0.319	0.322
	Burkina Faso	Lobi Bouroum	FLB	125	0.266	0.266
	Niger	Kouri Pur	NKO	35	0.337	0.333
	Mali	N’Dama	AND	45	0.281	0.280
West African Crossbreds (WAC)	Niger	Kouri Arabe	NKA	27	0.351	0.339
	Niger	Kouri x Bororo	NKB	34	0.344	0.338
	Benin	Bourgou x Zebu	EBZ	24	0.343	0.340
West African Zebu (WAZ)	Niger	Arabe Pur	NAR	48	0.340	0.341
	Niger	Bororo Diffa	NBD	21	0.348	0.340
	Niger	Bororo site Kouri	NBK	43	0.335	0.330
	Niger	Bororo Maaoua	NBM	25	0.335	0.331
	Benin	Goudali Benin	EGO	101	0.340	0.325
	Benin	Yakana	EYK	29	0.348	0.339
	Benin	Zebu Peuhl	EZP	71	0.327	0.339
	Burkina Faso	Bororo Burkina	FBO	14	0.354	0.342
	Burkina Faso	Goudali	FGO	43	0.335	0.336
	Burkina Faso	Zebu Peuhl Burkina	FZP	39	0.333	0.337
East African Zebu (EAZ)	Sudan	Arishy	ARI	25	0.344	0.336
	Sudan	Baggara	BAG	21	0.339	0.341
	Sudan	Butana x Kenana	SBK	20	0.364	0.338
	Sudan	Butana	BUT	41	0.340	0.328
	Sudan	Kenana	KEN	39	0.335	0.334
Southern African Zebu (SAZ)	Madagascar	Malagasy Zebu	MGZ	56	0.313	0.312
	Zambia	Angoni	ZAG	28	0.337	0.337
	Zambia	Baila	ZBI	50	0.333	0.334
	Zambia	Barotse	ZBR	24	0.328	0.327
	Zambia	Tonga	ZTG	51	0.335	0.332
	South Africa	Nguni	ANI	153	0.316	0.315
Southern African Crossbreds (SAC)	Madagascar	Manjani Boina	MJB	17	0.383	0.342
	Madagascar	Renitelo	REN	52	0.336	0.319
South Asian Zebu (SAZ)	India	Kangayam	IKA	37	0.312	0.300
	India	Nellore	ION	33	0.359	0.334
	Pakistan	Red Sindhi	PRS	36	0.318	0.310
	Pakistan	Sahiwal	PSH	38	0.322	0.31
	Pakistan	Tharparkar	PTP	30	0.321	0.321

The genome wide typing was done on an Affymetrix-Axiom platform using Bovine BovMDv3 array that consisted of 63,655 SNPs. The genotype data was extracted from .cel files following the Best Practices Workflow implemented in Axiom Analysis Suite version 5.1.1. Quality control was implemented in PLINK v2.0^26,27 in two parts with the first part done globally to remove SNPs with < 90% call rate, SNPs with less than 0.05 minor allele frequency (MAF) and samples with more than 5% missing genotypes. A total of 34,774 SNPs and 1965 cattle were left across the breeds. Hardy–Weinberg-Equilibrium (HWE) threshold of 10^–6 was applied separately for each breed group, with 34,744 SNPs remaining for West Africa taurine, 34,580 for African zebu, 31,961 for South Asian zebu, and 34,585 for European taurine. A total of 31,645 SNPs were common SNPs among the groups. These common SNPs were further pruned for high linkage disequilibrium using the parameter –indep 50 5 2 to reduce ascertainment bias and exclusion of closely related individuals using the parameter –king-cut-off 0.354 in PLINK v2.0²⁷. This process left 28,315 SNPs and 1898 cattle for subsequent analysis (populations structure). The second part of the quality control was done for subsequent analysis of effective population size (Ne) and runs of homozygosity (ROH) using –geno 0.1, –mind 0.05 and –king-cut-off 0.354 to remove SNPs with call rate less than 0.90, individuals with genotype call rate less than 95% and duplicate animals or twins from the data set using PLINK v2.0. A set of 44,190 SNPs and 1965 cattle were retained.

Genetic diversity and population structure

Genetic variability was measured using expected heterozygosity, F_ST and inbreeding coefficient. Expected heterozygosity was calculated on 28,315 SNP dataset using PLINK v2.0, inbreeding coefficient was calculated from runs of homozygosity using cgaTOH²⁸, and F_ST was calculated according to Weir and Cockerham²⁹ on 44,190 SNP dataset using dartR package in R³⁰. cgaTOH allows us to investigate ROH based inbreeding coefficients (F_ROH) by analyzing the proportion of ROH segments at different lengths (1–2 Mb, 2–4 Mb, 4–8 Mb, 8–16 Mb and > 16 Mb) within the total length of autosomes. Assuming the physical distance of 1 Mb being equivalent to the genetic distance of 1 centimorgan (cM), this corresponds with inbreeding to common ancestors back to 50, 25, 12.5, ~ 6 and ~ 3 generations in the past, respectively³¹. The following parameters were applied across ROH size classes: a minimum of 15 consecutive SNPs per ROH, a maximum physical gap of 1 Mb between adjacent SNPs, and variable thresholds for missing and heterozygous genotypes (maximum missing genotypes per window = 0 for 1–2 Mb and 2–4 Mb ROH; 1 for 4–8 Mb ROH; 2 for 8–16 Mb ROH; and 4 for > 16 Mb ROH; maximum heterozygous genotypes per window = 0 for all ROH classes except for ROH ≥ 16 Mb, which was kept as 1). The parameters were set using the -min_length, -max_gap, -max_missing, and -max_hetero flags, with clustering disabled (-skip_clustering) and the -force_proceed option enabled. Population structure of the breeds was investigated on 28,315 SNP dataset using three measures (admixture levels, principal components, and Reynolds distance). To estimate the admixture levels, we used ADMIXTURE software to determine ancestries assuming ancestral populations (K values) ranging from 2 to 30³² with at least 10 runs for each K. The software uses cross validation to find the optimal k value, the one with the lowest cross validation errors is assumed to be appropriate, k = 25 in this study (Supplementary Information S2 file). Although K = 25 showed minimal CV error indicative of best predictive fit under the ADMIXTURE model, lower K values (e.g., K = 2–10) were also used to interpret broad regional ancestry patterns, while higher K primarily captured finer population substructure. The resulting proportion of ancestry for each individual in each K was post-processed and visualized using pong. We also used Principal Component Analysis (PCA) using the –pca function in PLINK software, the resulting eigenvalues were visualized in R. Reynolds distance was calculated using a custom script and the distances were used to construct a neighbor-net tree using SPLITSTREE software³³. Gene flow and migrations were inferred using 28,315 SNP dataset on TREEMIX software. TreeMix was run with migration edges ranging from 0 to 8 (m = 0–8), with five independent replicates per edge and N’Dama (the West African taurine breed) as root population. Key parameters included the following: -k 1000 for SNP sampling blocks, -iter 100 for optimization iterations, -bootstrap 50 for branch length standard errors, -se 100 for migration weight standard errors, and -global for topology optimization across replicates³⁴.

Effective population size (Ne) and linkage disequilibrium

The historical trends and contemporary effective population size (Ne) were estimated using (GONE)³⁵ and CurrentNe³⁶ respectively on 44,190 SNP dataset. GONE estimates historical effective population size (Ne) by assuming that the observed linkage disequilibrium (LD) between pairs of loci separated by a genetic distance of approximately 1/(2t) Morgans reflects the Ne ‘t’ generations ago. This approach accounts for the cumulative effects of genetic drift and recombination over previous generations, as these forces shape the extent of LD observed at different genomic distances³⁵. On the other hand, CurrentNe accounts for different mating systems and small sample sizes while requiring less computational power³⁶. The method is based on artificial neural networks, and utilizes genome-wide SNP data, focusing specifically on loci in linkage disequilibrium (LD). Unlike earlier tools, CurrentNe provides estimates accompanied by confidence intervals, enhancing the robustness and interpretability of the results.

Results

Genetic diversity

Heterozygosity results (Table 1) showed Manjani Boina (MJB), a Southern African breed from Madagascar having the highest observed (0.383) and expected heterozygosity (0.354). The lowest heterozygosity was observed in Lagunaire (ELN) (0.261) and Lobi (FLB) (0.261), the two West African taurine breeds. Comparing the heterozygosity within the African taurine group, Bourgou (EBG) had the highest observed heterozygosity of 0.342. The estimated genomic inbreeding levels (F_ROH) were variable in the African populations. The estimates of inbreeding based on F_ROH was observed to be negligible with ROH segments of length 4 Mb and more (corresponding to 12 and less generations in the past) indicating limited recent inbreeding. The estimates of historical inbreeding based on short ROH segments of 1–2 Mb (F_ROH1) and 2–4 Mb (F_ROH2) length (Fig. 1) revealed relatively higher levels of inbreeding in West African taurine cattle. With the exception of Bourgou (EBG), Pabli (EPB), Kouri X Arabe (NKA) and Kouri (NKO), all the other West African taurine breeds showed > 2% estimates of F_ROH1 reflecting ancient inbreeding. The highest estimate of 5.8% F_ROH1 was observed in Lagunaire (ELN) cattle followed by N’Dama (AND) with 4.3% and Lobi with 4.0%. With respect to estimates of F_ROH2, > 1% level of inbreeding was observed in Lagunaire (ELN)cattle. With respect to West African zebu cattle, > 2% estimates of F_ROH1 were observed in Bororo Maaoua (NBM) while average F_ROH2 estimates were < 2% in all the other breeds from the region. The estimates of both F_ROH1 and F_ROH2 were < 2% in all the East African zebu cattle breeds. Similarly, the F_ROH2 estimates were < 2% in all the Southern African cattle while F_ROH1 estimates were < 2% in all but Nguni (ANI) from South Africa, Malagasy zebu (MGZ) and Renitelo (REN) from Madagascar and Barotse (ZBR) from Zambia. Genetic differentiation between breeds was investigated by F_ST, the results ranged from 0.01 to 0.39. Pairwise F_ST estimates were highest between African taurine (N’Dama-AND, Lobi-FLB and Lagunaire-ELN) and Asian zebu cattle breeds (Kangayam-IKA, Ongole-ION, Red Sindhi-PRS, Sahiwal-PSH and Tharparkar-PTP) ranging from 0.36 to 0.39. The pairwise estimates of F_ST between African taurine and African zebu breeds ranged between 0.13 and 0.29 (Fig. 2). Within the African zebu breeds, the highest recorded F_ST was 0.18 between Malagasy zebu (MGZ) and South African Nguni (ANI) cattle while the pairwise estimates among the rest of African zebu breeds were below 0.05. F_ST values among African taurine breeds were between 0.01 and 0.12.

Fig. 1.

Boxplot showing inbreeding coefficient (F_ROH) estimated from runs of homozygosity for African taurine and zebu populations from West, East and Southern Africa.

Fig. 2.

Population differentiation based on pairwise F_ST from 9 Western African taurine; 13 Western African zebu and crossbreds; 5 East African zebu; 8 Southern African zebu and crossbreds; and 5 South Asian zebu reference populations.

Population structure

Genetic relationship among West African taurine, West African zebu, East African zebu and Southern African zebu breeds, including European taurine and Asian zebu cattle as outgroups, was investigated and visualized using PCA (Fig. 3a). We considered the first and second principal components that explained 38.42% and 16.22% variation respectively. The results showed distinct and separate clustering of European taurine and Asian zebu cattle (Fig. 3a). Among the African cattle (Fig. 3b), three West African taurine breeds (Lagunaire-ELN, N’Dama-AND, and Lobi-FLB) clustered distinctly from the zebu cattle breeds indicating their level of ancestry proportion. However, other West African taurine populations such as Pabli (EPB), Gourounssi Tenado (FGT), Somba (ESO) and Gourounsii Nahouri (FGN) were observed to be scattered in between African zebu and African taurine clusters indicating significant levels of zebu admixture in them. Among the African zebu cattle, the East African cattle breeds from Sudan formed a distinct cluster. Similarly, Southern African zebu cattle breeds from Zambia (Baila-ZBI, Barotse-ZBR, Angoni-ZAG, Tonga-ZTG) and South Africa (Nguni-ANI) clustered together. The separate clustering of Renitelo (REN) and Manjani Boina (MJB) from Madagascar is understandable as both the breeds were developed by crossbreeding local Malagasy zebu (MGZ) with European taurine cattle. Interestingly, Malagasy zebu (MGZ) from Southern Africa did not cluster with the Southern African zebu group, instead clustered with the East African zebu group.

Fig. 3.

Principal component analysis for African taurine, African zebu populations with European taurine and South Asian zebu as reference populations. (Breed names Table 1).

Breed ancestry was estimated using the unsupervised Bayesian clustering approach as implemented in ADMIXTURE software. All the cattle in the dataset were clustered with models assuming K = 2 to K = 25 (Fig. 4a). The admixture results were consistent with PCA, at K = 2 there was a clear separation of taurine and indicine ancestry among the breeds or populations. The European taurine (FLV, IJR, SHO, YBS) and Asian indicine (IKA, ION, PRS, PSH, PTP) reference populations revealed > 97–99% proportions in their respective ancestry. The African zebu and crossbred cattle showed varying levels of indicine and taurine ancestry in them (Fig. 4b). Amongst the African taurine breeds, Lobi (FLB) and N’Dama (AND) showed 6.9% and 5.7% zebu ancestry, respectively while Lagunaire (ELN) showed a slightly higher level of 8.8% zebu ancestry. With respect to other taurine breeds (Bourgou-EBG, Somba-ESO, Pabli-EPB, Kouri Pur-NKO, Gourounsii Nahouri-FGN, Gourounsii Tenado-FGT), zebu introgression ranging from 5 to 48% was observed. In African zebu cattle breeds, the proportion of taurine-indicine admixture varied across different regions. The proportion of indicine ancestry was highest among the East African zebu breeds such as Arishy (ARI), Baggara (BAG), Butana × Kenana (SBK), Butana (BUT) and Kenana (KEN) ranging from 68 to 78%, followed by West African zebu cattle. The lowest proportion of indicine ancestry ranging from 53 to 67% was observed among the Southern African cattle (Angoni-ZAG, Baila-ZBI, Barotse-ZBR, Tonga-ZTG and Nguni-ANI) with the exception of Malagasy zebu (MGZ) cattle located in the Indian ocean island country of Madagascar. At K = 3, the European and African taurine clustered distinctly providing better clarity on zebu-taurine admixture among African cattle. At K = 4, zebu ancestry within African cattle showed two distinct clusters. The evidence for admixture of zebu genes was more pronounced in the East and West African zebu as compared to Southern African zebu cattle indicating the ongoing dynamics of zebu-taurine admixture in these regions. Overall, the African zebu cattle breeds showed a wide range of Asian zebu ancestry. Beginning K = 6, Malagasy zebu clustered separately from the rest of African zebu cattle, potentially owing to their reproductive isolation. Similarly, K = 6 to K = 10 showed Goudali (EGO), the West African zebu to be genetically distinct from other indicine breeds in the region. Neighbor-net tree constructed using Reynolds distance was consistent with the population structure results from PCA and admixture analysis. East African and Southern African zebu breeds were clustered together whilst West African taurine and the majority of West African zebu branched separately from each other (Fig. 4c).

Fig. 4.

(a) Admixture plot of African zebu and taurine populations from West, East and Southern Africa (b) Neighbour-net tree of 44 populations based on Reynolds distance matrix. (Breed abbreviations Table 1).

Effective population size

The effective population size (Ne) provides crucial insights into the demographic history of cattle breeds, reflecting past population contractions, expansions, and genetic drift. In this study, historical Ne was reconstructed over the past 700 generations using GONE software, with a focus on the most recent 100 generations for visualization. Contemporary Ne was estimated using CurrentNe to infer ongoing population dynamics. The results revealed substantial heterogeneity in Ne trends across African cattle breeds, shaped by distinct historical events, genetic bottlenecks, and breed management strategies. West African cattle exhibited marked differences in Ne trends between taurine and zebu populations. Among the West African taurine breeds, N’Dama (AND), Lobi Bouroum (FLB), Somba (ESO), and Gourounssi Tenado (FGT) displayed relatively stable historical Ne trends, with a notable reduction occurring approximately 65–80 generations ago. Kouri (NKO) and Gourounssi Nahouri (FGN) deviated from this pattern, displaying a more gradual decline without a sharp bottleneck, suggesting a relatively larger and more stable ancestral population. In contrast, the West African zebu breeds exhibited pronounced fluctuations in Ne. The Nigerien zebu cattle breeds, Arabe Pur (NAR), Bororo Diffa (NBD), Bororo Maaoua (NBM), Bororo site Kouri (NBK), Bororo x Kouri crossbred (NKB) and Kouri × Arabe Pur crossbred (NKA) experienced severe bottlenecks approximately 60–90 generations ago, followed by a phase of demographic recovery. The zebu cattle breeds of Burkina Faso (Bororo-FBO, Goudali-EGO and Zebu Peuhl-EZP) and Benin (Yakana-EYK) displayed similar population bottleneck events between 40 and 70 generations before present, however the subsequent demographic recovery was not significant. The Zebu Peuhl (EZP) of Benin underwent an early bottleneck followed by steady expansion, suggesting resilience against genetic erosion.

East African cattle, predominantly zebu, exhibited substantial variation in Ne trajectories with Kenana experiencing significant bottleneck 35 generations ago. Butana (BUT) maintained the historical Ne, suggesting long-term population stability, possibly due to their historical role as dominant breeds in Sudanese pastoralist systems. In contrast, Arishy (ARI) and Baggara (BAG) demonstrated moderate historical Ne values with greater fluctuations, reflecting the impact of localized breeding practices and genetic introgression events. Southern African zebu populations showed distinct Ne trends compared to their East and West African counterparts. Angoni cattle exhibited the highest historical Ne among Southern African breeds, with fluctuations between generations 100 and 40, followed by a sharp decline at generation 37. Barotse cattle displayed a biphasic trend, with an initial period of stability followed by a steep Ne decline beginning around 50 generations ago. Nguni (South Africa), Tonga, and Baila (Zambia) exhibited the lowest historical Ne among Southern African breeds between generations 100 and 60, followed by a pronounced decline. The sharpest reductions were observed in Tonga, Nguni, and Baila, reflecting the impact of genetic drift and restricted gene flow in these populations. Malagasy zebu exhibited a distinct Ne trajectory, with a sharp decline approximately 65 generations ago that persisted until generation 55.

Contemporary estimates of Ne, inferred using CurrentNe, revealed further insights into ongoing population dynamics. Broadly, the contemporary Ne estimates ranged from 17 to 391 and were largely concordant with the historical demographic patterns inferred by GONE. Seven breeds, including two taurine (Lagunaire (ELN), Gourounssi Nahouri (FGN)), and five zebu types (Zebu Peuhl (EZP), and Butana (BUT), Butana × Kenana (SBK), Manjani Boina (MJB), and Malagasy Zebu (MGZ)) exhibited contemporary Ne estimates below 50, suggesting potential risks of inbreeding and reduced genetic diversity.

Gene flow and migrations

Treemix analysis revealed a complex history of gene flow and admixture among African cattle populations, indicating multiple migration events shaping the genetic structure of taurine and zebu breeds. The maximum likelihood tree reconstructs population splits and drift patterns, while the presence of migration edges highlights historical interbreeding events. Migration events were sequentially modelled up to eight migration edges to capture the most probable historical introgression signals, the optimal topology was observed with four migration edges (m = 4) (Fig. 5). Five South Asian zebu breeds viz. Sahiwal (PSH), Red Sindhi (PRS), Tharparkar (PTP), Ongole (ION), and Kangayam (IKA) were incorporated as reference populations. The observed migration edges, weighted by their intensity, provided evidence of the introgression dynamics among East African, West African, and Southern African zebu, as well as West African taurine breeds. The genetic structure of East African zebu, represented by Kenana (KEN), Baggara (BAG), and Butana (BUT) breeds, showed significant indicine introgression. The Treemix analysis revealed strong migration edges from South Asian indicine cattle into these breeds, consistent with historical records of zebu introduction through the Horn of Africa. Conversely, taurine breeds such as N’Dama (AND), Lobi Bouroum (FLB), and Somba (ESO) exhibited minimal gene flow from zebu populations. Southern African zebu breeds, including Nguni, Baila, and Barotse, displayed distinct genetic signals compared to East and West African zebu.

Fig. 5.

Treemix analysis African taurine and zebu populations including South Asian zebu as reference using a model assuming k = 1000 and up to 8 migrations. The optimal topology was observed with four migration edges (m = 4).

Discussion

Genetic diversity and inbreeding

The assessment of genetic diversity within African cattle populations is a critical step in understanding their evolutionary history, adaptability and their utility in sustainable livestock development^10,37. The results from the present study corroborate prior observations that African cattle exhibit remarkably high levels of genetic diversity, particularly when compared to European taurine and Asian indicine breeds^37,38. Further, the lower genetic diversity of West African taurine breeds such as Lagunaire (ELN) and N’Dama (AND), compared to African zebu breeds, aligns with previous findings indicating a narrower genetic base in these trypanotolerant cattle ⁴. The plausible reasons for low genetic diversity in West African taurine cattle may include (i) the legacy of an initial low effective population size (ii) population bottleneck following disease challenges and (iii) prolonged periods of genetic isolation in tsetse-infested zones that limited the gene flow and introgression from zebu. This is supported by the fact that relatively higher genomic diversity observed in Bourgou (EBG) and Pabli (EPB) where the prevalence of tsetse and associated trypanosomiasis is not higher. The reduced pressure of tsetse/trypanosomiasis in their native tracts did not hinder the gene flow particularly from zebu cattle into these breeds.

In contrast, the elevated heterozygosity levels observed in zebu and admixed populations, such as the Malagasy zebu (MGZ), Manjani Boina (MJB) and West African crossbred types (e.g., Kouri-Arabe (NKA) and Bourgou-zebu (EBZ)), reflect extensive historical and ongoing gene flow from indicine cattle, likely a result of centuries of trans-Saharan and Indian Ocean trade routes introducing Bos indicus genes into African cattle populations¹⁹. The high diversity estimates also suggest substantial intra-population genetic variability, particularly within zebu and admixed populations, and are consistent with earlier observations of heterozygosity in African cattle across different ecological zones and breed groups³⁷. The present study also highlights significantly high heterozygosity among East African and Sahelian zebu breeds. This aligns with earlier work on Cameroonian Goudali and Ethiopian highland cattle, which reported similarly elevated heterozygosity values, suggesting long-term maintenance of variation through population admixture and low-intensity selection pressures^21,24. Furthermore, South African Nguni cattle, and other East African zebu breeds (e.g., Ogaden, Boran), have also been characterized by extensive heterozygosity and low recent inbreeding³⁹. These breeds benefit from complex demographic histories involving admixture with both taurine and zebu lineages, adaptive selection for thermal stress tolerance, and disease resistance.

Inbreeding assessed using runs of homozygosity (F_ROH) provides insight into population isolation and the extent of recent common ancestry. The present study revealed that African cattle generally exhibited lower levels of inbreeding compared to their European and Asian counterparts. This is exemplified by relatively short and infrequent ROH segments across most African breeds, consistent with previous findings^40,41. Although medium-density SNP arrays may underestimate short ROHs, the observed pattern dominated by short segments with few ROH > 4 Mb is consistent with WGS-derived estimates reported for African cattle ²². African Taurine breeds such as Lagunaire (ELN), Lobi (FLB) and Bourgou (EBG) showed elevated values of F_ROH indicative of relatively higher homozygosity. This likely reflects long-term isolation, small effective population sizes, and possibly intensive selective pressure. The relatively high F_ROH levels observed in some African taurine populations can also be attributed to farmers aiming to keep their animals purebred due to high prevalence of trypanosomiasis⁴². In contrast the low F_ROH levels in the indicine breeds could be due to extensive gene flow and dynamic zebu introgression commonly observed in most African cattle populations⁴³. For example, East African zebu breeds exhibit uniformly lower F_ROH values. Both F_ROH1 and F_ROH2 values remained less than 0.03, with minimal variation among breeds. West African zebu breeds fall intermediate between African taurine and East African zebu, with moderate F_ROH1 values. The breed Arabe Pur (NAR) from Niger showed a relatively high F_ROH1, potentially indicating localized inbreeding. Southern Africa zebu breeds also exhibit moderate F_ROH levels. Notably, Nguni (ANI) and Malagasy zebu (MGZ) cattle displayed higher F_ROH1 values compared to other regional breeds, which may reflect demographic contractions or recent founder effects^44,45. Most small-holder farmers in Africa share communal grazing, water points and young bulls hence providing the possibility of cross-herd mating which lowers the inbreeding coefficients⁴⁵. These findings collectively highlight the heterogeneity of inbreeding levels across African cattle, driven by differential management practices, breeding systems, and historical population dynamics. The relatively low inbreeding in most populations signals resilience and suggests that genetic erosion remains limited, although population monitoring is necessary for more isolated breeds. The F_ST between the African taurine breeds (0.01–0.12) in this study was relatively higher than those reported in previous studies (0.02 for Lagunaire (ELN) vs. Somba (ESO), 0.05 for Lagunaire (ELN) vs. Bourgou (EBG), and 0.03 for Somba (ESO) vs. Bourgou (EBG))⁴⁶. The higher values (F_ST > 0.08) were mainly due to higher pairwise F_ST between Bourgou and N’dama, Bourgou and Lagunaire, Bourgou and Lobi while the pairwise F_ST values among other taurine breed pairs ranged from 0.01 to 0.08. The relatively higher differentiation observed in this study likely reflects substantial zebu introgression in populations such as Bourgou and Pabli, which has increased their genetic divergence from the other taurine breeds. Between indicine breeds the pairwise difference was generally less than 0.05⁴⁴. However, it is important to note that the medium-density SNP array used for F_ROH estimation may underestimate short ROH segments and bias the fine scale differentiation due to limited marker resolution; therefore, these results need to be interpreted with caution.

Population structure and admixture

The genomic structure of African cattle reflects a complex history of domestication, migration, admixture, and adaptation to diverse ecological pressures across the continent¹⁰. The African cattle genome is a mosaic of taurine and indicine ancestry with varying levels of zebu introgression across different regions^37,48. The earliest domestic cattle in Africa were described as humpless taurine (Bos taurus), introduced approximately 7000 to 8,000 years before present (BC) from the Fertile Crescent, potentially admixed with wild African aurochs (Bos primigenius mauritanicus) during their initial dispersal into North and West Africa^10,19,47. These cattle adapted to the humid, sub-humid, and tsetse-infested ecological zones of West Africa, where they remain predominant in the form of trypanotolerant taurine breeds such as N’Dama (AND), Lagunaire (ELN), and Lobi (FLB). Subsequent introgression of zebu cattle, particularly through the Horn of Africa around 2000–2500 BC and later during the Islamic expansion along the East African coast, has been reported to have significantly reshaped the genomic landscape of African cattle^10,24. The results of the ADMIXTURE analysis at K = 2 and higher resolutions reveal significant variation in zebu introgression among African cattle, influenced by both geography and environmental selection pressures. These findings are consistent with the historical and genetic reconstructions of African pastoralism and cattle diversification reported in earlier studies^4,10,37.

The results of the present study demonstrate the persistence of high taurine ancestry in certain West African breeds. The low levels of zebu admixture in N’Dama (5.7%), Lagunaire (8.8%) and Lobi (6.9%) reaffirm their identity among the few remaining pure African taurine populations. These findings are in agreement with earlier genome-wide analyses that highlighted the unique trypanotolerance of these breeds, attributed to selective pressures from endemic trypanosomiasis vectored by the tsetse fly (Glossina spp.)^4,37,48. The high frequency of genomic regions under selection related to immune response and anemia resistance in N’Dama supports this adaptive landscape⁴⁹. Conversely, other West African taurine breeds such as Somba (ESO), Gourounssi (FGN and FGT), and Kouri Pur (NKO) have significantly higher indicine ancestry, ranging from 22% to over 44%. These variations likely reflect historical admixture with zebu cattle following Fulani pastoralist expansions, as documented in both molecular and ethnographic studies. Fulani pastoralism, involving transhumance movements across the Sahel and Savannah belt, has facilitated zebu introgression into taurine populations, especially in regions with lower tsetse challenge where trypanotolerance characteristics are less critical to survival^19,24.

The pattern of indicine ancestry is more pronounced in East African zebu populations. Breeds such as Butana (BUT), Kenana (KEN), and Baggara (BAG) showed > 68% zebu ancestry, with Butana × Kenana (SBK) composites reaching 77%. These values aligned with earlier findings that Horn of Africa and East Africa served as a primary entry point for zebu cattle into the continent, where indicine introgression became widespread due to both ecological compatibility and historical trade and migration routes^19,21. The relatively arid climate of East Africa, combined with the utility of zebu cattle under high temperature, parasite burden, and feed-limited conditions, likely promoted their rapid diffusion⁴⁹.

Southern African cattle displayed a broader spectrum of zebu ancestry. Notably, Nguni (ANI) cattle, central to traditional livestock production in southern regions, retained a higher taurine background (67%) compared to other Southern African breeds such as Barotse (ZBR) and Tonga (ZTG). This limited zebu introgression likely reflects the delayed introduction of zebu in the region, lower historical zebu influxes. In contrast, Malagasy zebu (MGZ) from Madagascar showed high zebu ancestry (77%), consistent with a more recent and isolated introduction of indicine cattle into the island’s agroecosystem. The level of zebu admixture in Malagasy zebu (MGZ) is comparable to that of East African zebu breeds, further providing evidence to the hypothesis that zebu immigration in the continent might have followed through local Arabian contacts or part of the long-distance Indian Ocean trade route^10,50. From K = 6 onwards, the ADMIXTURE analysis delineated a distinct cluster associated with Southern African cattle, particularly the Nguni (ANI) breed. This differentiation becomes increasingly pronounced at higher K values, distinguishing Nguni from both East and West African zebu-taurine composites and from other Southern African breeds. Makina et al.⁵¹ described the Nguni as a classical Sanga breed, characterized by a predominantly African taurine genome (~ 60%), with moderate indicine (~ 30%) and minor European taurine (~ 10%) introgression. The Nguni cattle exhibited a unique admixture signature, distinct from the Zambian breeds (Angoni -ZAG, Baila -ZBI, Barotse -ZBR and Tonga -ZTG) that showed relatively higher indicine influence and reduced taurine ancestry. This finding is consistent with previous reports that showed Nguni and Afrikaner cattle as primary recipients of African taurine ancestry, with lower indicine admixture relative to other Southern African breeds⁵¹.

The bifurcation of the indicine cluster at K = 4 is particularly revealing, as it likely reflects the dual origin of zebu ancestry in African cattle. A pronounced bifurcation within the zebu component was observed in East and West African versus Southern African zebu cattle populations. This pattern indicates that East and West African zebu cattle harbor indicine ancestries different from Southern African zebu, derived from at least two genetically distinct indicine source populations. Such duality likely reflects temporally and geographically separate introgression events and routes of zebu introduction into the African continent. In East Africa, the indicine ancestry appears predominantly derived from South Asian zebu, particularly from the breeds located near the domestication site of zebu in Indus valley region represented by modern day Eastern parts of Pakistan and Northwestern parts of India. This was also evident in the genetic profile of East African breeds such as the Ethiopian Boran, Kenana, and Butana, which show strong affinity to Indian zebu breeds. Historical records and archaeological data support this scenario, indicating the arrival of South Asian zebu into East Africa via maritime trade routes in the Indian Ocean during the first millennium AD, most intensively after 700 AD^11,52. Genome-wide modelling by Kim et al. ⁴⁷ using qpAdm and qpGraph approaches revealed that while most African zebu populations can be explained by admixture with North Indian indicine, the residual ancestry patterns suggest additional affinities with Southeast Asian indicine sources, such as the Jian breed from South China. Although minor (typically 0.8%–6.5%), these contributions were statistically significant and consistently detected across multiple analytical frameworks⁴⁷. The failure of models relying solely on North Indian indicine cattle to explain African zebu supports the hypothesis of a complex, polyphyletic indicine ancestry. Therefore, the bifurcation in indicine ancestry at K = 4 does not merely represent regional drift but likely captures a fundamental difference in source populations and historical introgression pathways⁴⁰.

Effective population size, gene flow and migrations

The Ne trajectories reconstructed in African cattle using GONE provide crucial insights into their demographic history and the historical pressures that have shaped their current genetic structure. One notable finding from these analyses is the presence of bottleneck signals in several African cattle populations between 35 and 90 generations ago, a pattern suggestive of continent-wide disturbances. When assuming a generation interval of 4.5 years in cattle ^53–55, this time window corresponds approximately to the late 19th and early twentieth centuries a period that aligns with the devastating pan-African rinderpest epizootics between 1887 and 1905. The inferred Ne bottlenecks is consistent with the timing of rinderpest outbreak in the region. The rinderpest virus, a morbillivirus of high virulence, swept across Africa after its initial introduction in Eritrea in 1887, resulting in mortality rates exceeding 90% in susceptible cattle populations⁵⁶. The disease rapidly traversed the continent, reaching southern Africa by 1896, where it induced catastrophic herd losses^57,58. These events decimated both indigenous livestock and wildlife, particularly ruminants, leading to the collapse of pastoral economies and substantial loss of genetic diversity in affected populations. The GONE-based Ne estimates mirror these historical impacts. For instance, the bottlenecks observed in West African zebu populations (e.g., Bororo) (Fig. 6a) and Southern African breeds (e.g., Nguni (ANI), Tonga (ZTG), and Barotse (ZBR)) (Fig. 6b) are temporally consistent with the period of rinderpest incursion. These signals are likely to reflect both the direct mortality imposed by the virus and secondary demographic consequences such as reduced reproduction, abandonment of herds, and shifts in management practices. Moreover, East African breeds such as Kenana (KEN) exhibit pronounced bottlenecks at approximately 35 generations ago (~ 160 years) (Fig. 6c), suggesting local intensification of the epidemic or subsequent population instability post-outbreak. While other factors such as droughts, disease co-morbidities, or socio-political upheavals might also have contributed to demographic contractions, the wide geographic scope and temporal alignment with rinderpest point to it as a possible driving force. This is supported by previous genomic analyses which similarly noted effective population size reductions in African cattle corresponding to known historical pathogen pressures³⁷. Such perturbations can leave enduring signatures that influence current genetic diversity and adaptive potential. However, the inference of historical bottlenecks must be interpreted with consideration of methodological constraints. The GONE algorithm robustly identifies signals of past reductions in effective population size (Ne) but cannot establish their specific cause. Therefore, while the rinderpest outbreaks represent a plausible explanation, our results should be viewed as supporting a strong temporal association, with causation inferred from the historical record rather than proven by the genomic data alone.

Fig. 6.

Effective population size trends estimated using GONE software for taurine and zebu populations from (a) West Africa (b)Southern Africa (c) East Africa (d) Ne estimated by integration over the whole genome using CurrentNe software. (Breed abbreviation Table1).

Contemporary estimates of effective population size derived from CurrentNe offer vital insights into the ongoing genetic health of African cattle populations (Fig. 6d). Small Ne values observed in several breeds are particularly concerning, as they elevate the risk of inbreeding depression, loss of adaptive alleles, and reduced capacity to respond to future environmental and disease challenges⁵⁹. Breeds such as Lagunaire (ELN), Gourounssi Nahouri (FGN), Zebu Peuhl (EZP), and Butana (BUT) showed Ne values below the critical threshold of 50. These findings reflect not only the legacy of historical bottlenecks but also contemporary forces such as reproductive isolation, artificial selection, and shrinking breeding populations, particularly in smallholder and pastoral systems⁶⁰. The discrepancy in Ne between breeds also underscores varying levels of gene flow, management intensity, and population connectivity. Breeds with higher contemporary Ne such as Angoni (ZAG) or Malagasy zebu (MGZ) may benefit from greater demographic buffering, genetic heterogeneity, or integration with broader metapopulations. Conservation and breeding programs should thus prioritize genomic monitoring of low Ne populations, encourage sustainable breeding structures, and promote strategies that maintain or increase Ne to safeguard genetic diversity and adaptability.

The complex gene flow patterns among African cattle populations, reflect historical introgression events that shaped the genetic architecture of African taurine and zebu breeds. Admixture between African taurine and zebu cattle is particularly prominent across the Sahelian and East African regions, consistent with earlier reports suggesting recurrent introgression events following the introduction of Asian zebu into the continent^40,61. The observed gene flow gradients suggest the consistent replacement of taurine ancestry with that of zebu in response to increasing aridity and parasite burdens, which are known to influence host genetic composition^10,37. The differentiation between Indian-derived zebu and African zebu indicates that, despite shared ancestry, African zebu has undergone significant divergence likely due to local adaptation and further crossbreeding with indigenous taurine populations. This is consistent with genome-wide local ancestry studies showing partial retention of taurine loci associated with adaptive traits, such as trypanotolerance and thermotolerance^61,62. The observed gene flow further supports the hypothesis of a repeated and dynamic contact between imported Indian zebu and pre-existing African populations, likely facilitated by trade and pastoralist expansions⁶³. Similarly, the introgression from European and Middle Eastern taurine cattle into African taurine breeds, although less extensive, carries significant evolutionary implications. Importantly, while gene flow introduces beneficial alleles into recipient populations, it also raises concerns about the erosion of unique local adaptations. For instance, widespread zebu introgression into African taurine populations has been associated with the dilution of traits like trypanotolerance in breeds such as N’Dama and Baoulé¹⁰. This underlines the need for conservation strategies that balance improvement with preservation of unique adaptability, fitness and disease resistance traits in indigenous populations.

Conclusion

The present study is a comprehensive genome-wide assessment of African taurine and zebu cattle across three major regions. The study highlights the rich genetic mosaic that characterizes African cattle, shaped by a complex interplay of ancient domestication events, transcontinental migrations, and continuous gene flow in dynamic pastoral and agropastoral systems. An important conclusion of this study is that all breeds called African zebu in this study show at least 22% taurine admixture. African pastoral systems, characterized by mobility, low external inputs, and adaptation to highly variable environments, have applied a unique form of selection on cattle populations. Traits such as thermotolerance, drought resistance, disease tolerance, and low maintenance costs have been favored. The results confirmed significant regional heterogeneity in admixture levels, with West African taurine cattle retaining a predominantly taurine ancestry due to natural selection pressures operating in trypanosome-endemic zones, while East African and Malagasy zebu cattle possess the highest indicine ancestry proportions. Patterns of admixture and differentiation indicate at least two temporally and geographically distinct waves of indicine introgression, which contributed to the genomic heterogeneity seen in modern African zebu. Southern African breeds, particularly Nguni, retain a unique Sanga ancestry profile characterized by mixed taurine and indicine components with minimal European influence. Notably, our results confirm the reproductive isolation of Malagasy zebu populations and their close genetic affinity to South Asian zebu, supporting historical narratives of Indian Ocean trade-mediated cattle introductions. Estimates of effective population size (Ne) show that many African cattle populations experienced demographic contractions corresponding to the nineteenth century rinderpest epizootic. Contemporary Ne estimates are low in several taurine and localized zebu populations, suggesting vulnerability to inbreeding depression and loss of adaptive genetic variation. The evidence of declining Ne and high levels of admixture in specific populations underscores the urgency of implementing sustainable breeding strategies that preserve locally adapted genomes. Levels of inbreeding towards recent common ancestors up to 50 generations ago, are generally very low, most likely due to the nature of mating structures in communal systems. Collectively, our findings highlight the genetic richness and regional uniqueness of African cattle. The continued introgression and mixing of taurine-indicine gene pools, underscore the importance of genomic surveillance and conservation strategies that integrate historical admixture patterns, effective population management, and adaptation traits in African cattle. Although the present study provided significant insights and enhanced understanding of genomic diversity in African cattle, certain limitations need to be acknowledged. The use of medium-density SNP array data, while appropriate for population-level analyses, offers lower resolution than whole-genome resequencing (WGS). Unlike SNP arrays, WGS can capture rare and population-specific variants, as well as copy-number and structural variants that may potentially contribute to local adaptation or recent selection. Future incorporation of WGS data will be valuable for refining variant-level inferences and for more comprehensively assessing adaptive genomic variation within these underrepresented cattle population.

Author contributions

Conceptualization, K-P, J-S and G-M; Methodology TK-M and R-P; Breed Survey and Investigation; L-Z, A-T, ASR-T, H-H, Y-AH, B-C, N-R, T-H, S-R, C-B; Formal analysis, TK-M; Data curation, TK-M and K-P; Writing-original draft preparation, TK-M; Review and editing K-P, J-S and G-M; Project administration, K-P; Funding acquisition, K-P. All authors have read and agreed to the published version of the manuscript.

Funding

This project was funded under the IAEA coordinated research projects D31028 and D31030.

Data availability

All relevant data has been stated in the manuscript. The original SNP genotype data has been provided as Supplementary Information S3 file.

Declarations

Competing interests

The authors declare no competing interests.