Breed-Specific Genetic Recombination Analysis in South African Bonsmara and Nguni Cattle Using Genomic Data

Abstract

South African cattle comprise diverse breeds with distinct evolutionary histories, potentially reflecting differences in recombination landscapes. This study assessed genome-wide recombination rates and hotspots in Bonsmara (n = 190) and Nguni (n = 119) cattle using three-generation half-sib pedigrees genotyped with the Illumina Bovine SNP50 BeadChip. Phasing across 29 autosomes was conducted using SHAPEIT v2, and crossover events were inferred using the DuoHMM algorithm. The total number of crossover events detected was higher in Nguni (n = 8982) than in Bonsmara (n = 7462); however, the average recombination rate per 1 Mb window was significantly higher in Bonsmara (0.31) compared to Nguni (0.18) (p < 0.01). This apparent discrepancy reflects differences in genomic distribution and crossover clustering across breeds, rather than overall recombination frequency. A critical limitation of the study is the reliance on half-sib families with small family sizes, which may underestimate recombination rates due to limited meiotic sampling and increased variance in crossover detection. We identified 407 recombination hotspots in Bonsmara and 179 in Nguni, defined as intervals exceeding 2.5 standard deviations above the mean recombination rate. Genes such as PDE1B and FP which are associated with productions traits were located within hotspot-enriched regions. However, functional causality between these genes and local recombination activity remains unverified. Our results provide statistically supported evidence for breed-specific recombination patterns and hotspot distributions, underscoring the importance of incorporating recombination architecture into genetic improvement strategies for South African cattle.

1. Introduction

The South African beef cattle industry comprises a range of breeds that reflect the country’s diverse agro-ecological zones and adaptive requirements. These include indigenous breeds such as Nguni and synthetic breeds like Bonsmara, both of which are well regarded for their adaptability, fertility, and production potential under harsh and variable climatic conditions [1,2]. Understanding the genetic mechanisms that underpin these traits is critical for sustainable improvement programmes, especially under the increasing pressures of climate change and emerging disease threats. Among the five primary forces of evolution—mutation, migration, selection, genetic drift, and recombination—recombination remains relatively understudied in African cattle populations, despite its central role in shaping genomic variation and driving evolutionary processes [3].

Meiotic recombination involves the formation of crossover events between homologous chromosomes, resulting in novel allele combinations in gametes. This process is essential not only for appropriate chromosomal segregation, but also for generating genetic diversity that contributes to phenotypic traits such as adaptability, fertility, and production [4]. Recombination events are unevenly distributed along the genome, with high-frequency sites often clustering in narrow regions of 1–2 kilobases known as recombination hotspots [5,6,7,8]. These hotspots coincide with DNA double-strand breaks (DSBs), a prerequisite for crossover formation [5,9] and thus represent important foci for understanding genetic variability and inheritance.

The recombination landscape is influenced by several biological and environmental factors, including age, sex, chromosomal architecture, and regulatory proteins such as the PR domain containing 9 PRDM9 [3]. Recombination rates are also highly variable between species and populations, as documented in humans, mice, and birds [10]. Consequently, elucidating recombination dynamics across cattle breeds offers valuable insight into the genomic architecture of economically important traits and the evolutionary forces shaping them. Such insights are foundational for enhancing the resolution of genome-wide association studies (GWAS), improving linkage disequilibrium (LD) mapping, and detecting regions under recent selection [11,12,13].

Recent advances in genomic technologies have enabled a deeper understanding of African cattle breeds’ genetic composition and diversity [14]. These studies have revealed high levels of genomic diversity in African cattle, largely attributed to historical admixture, gene flow, selection, and drift [14,15]. However, while these forces contribute to genome-wide diversity, recombination serves as a primary source of intragenomic variation and structural reshaping [16]. Despite its centrality, few studies have systematically explored recombination in African beef breeds, and none have characterised recombination hotspot architecture in South African cattle [17,18]. This limits the ability to accurately map traits of interest and fully exploit local breeds’ genetic potential in selection schemes.

Several methods are available to infer recombination, including LD-based, gamete-based, and pedigree-based approaches [19,20,21]. Among these, pedigree-based methods offer the advantage of enabling sex-specific analyses and are less affected by historical demographic events [22,23]. However, the accuracy of recombination estimates from pedigree data is highly dependent on family size and structure. In particular, the use of half-sib families with small progeny sizes introduces limitations in recombination inference due to reduced numbers of observable meiosis and increased stochastic variance [11,24]. These constraints necessitate cautious interpretation of recombination metrics derived from such designs.

Given the scarcity of recombination studies in African cattle and the importance of recombination for both evolutionary and applied genetics, this study aimed to investigate recombination rates and identify recombination hotspots in two South African cattle breeds—Bonsmara and Nguni—using SNP genotype data and half-sib family structures by making use of pedigree-based methods. By contrasting these two breeds with divergent histories and phenotypic characteristics, this work seeks to characterise the extent of recombination variability and inform breeding strategies grounded in breed-specific genomic architecture.

2. Materials and Methods

2.1. Sampling and DNA Isolation

A total of 190 Bonsmara and 119 Nguni hair samples consisting of half-siblings were collected at the Biobank of the Agricultural Research Council, Animal Production and the Agricultural Research Council-Ncera farm located in the Eastern Cape Province. Hair samples were sent to the Agricultural Research Council, Biotechnology Platform (ARC-BTP) located in Pretoria, South Africa, laboratory for processing.

Genomic DNA was extracted from approximately 20 hair root follicles per individual. The hair roots were incubated overnight at 65 °C in Sodium Chloride-Tris-EDTA (STE) buffer with 10 µL Sodium Dodecyl Sulphate (SDS) and 30 µL Proteinase K (PK). The lysate was purified with Phenol–Chloroform–Isoamyl Alcohol (PCI), followed by ethanol precipitation at −20 °C. DNA pellets were air-dried and resuspended in 50 µL of nuclease-free distilled water. DNA concentrations were quantified using a Qubit^® 3.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s protocol. The DNA concentrations ranged from 10.18 ng/µL to 98.57 ng/µL. Samples were not normalised before genotyping to preserve the natural variance in DNA quality and yield.

2.2. SNP Genotyping

Genomic DNA from Bonsmara and Nguni cattle was genotyped using the Illumina BovineSNP50 BeadChip (Illumina Inc., San Diego, CA, USA), following the manufacturer’s protocol. The genotyping process involved fragmentation of DNA samples, precipitation, and resuspension in a hybridisation buffer, followed by hybridisation to the BeadChip. Single-base extension, multi-layer staining, and washing steps were conducted using proprietary Illumina reagents (RA1, PM1, FMS, TEM, XC1–XC4), after which the BeadChip was scanned using the Illumina iScan platform. The resulting data were exported in binary format for downstream genomic analysis. Detailed procedural steps are available from the manufacturer (https://www.illumina.com, accessed on 5 March 2024).

2.3. Data Filtering and Quality Control

Quality control on genotype data was performed with the use of PLINK v1.9 [25]. Genome-wide Mendelian inconsistencies were tested on all samples, and individuals with elevated amounts of uncalled genotypes that were higher than 95% were removed from the analysis. Individual SNPs with low call rates (CR < 0.95), low minor allele frequencies (MAF < 0.05), and a p-value for Hardy–Weinberg equilibrium test less than 0.0001 were filtered for further analysis. SNPs that had duplicate genomic positions were also excluded using PLINK v1.9. After quality control, the Bonsmara cattle breed had a remainder of 189 animals and 45,402 SNPs, while the Nguni cattle breed had 119 animals and 46,124 SNPs.

Family structure was reconstructed using both pedigree records and SNP-based parentage verification. In the Bonsmara population, 43 half-sib families were identified, with family sizes ranging from 3 to 9 (mean = 4.4; SD = 2.1). The Nguni dataset comprised 28 half-sib families, with family sizes ranging from 2 to 7 (mean = 4.3; SD = 1.8). This yielded 784 and 473 informative meioses in Bonsmara and Nguni, respectively. These structures were suitable for recombination mapping using pedigree-informed phasing approaches.

2.4. Haplotype Phasing

Haplotype phasing was performed for each of the 29 autosomes using SHAPEIT2 v2 [26]. The software models haplotypes in structured or admixed populations and is well suited for South African cattle breeds known for their composite genetic structure [3]. The phasing process involved 20 iterations of the Markov Chain Monte Carlo (MCMC) algorithm and sampled 20 haplotype pairs per individual per iteration [17]. To minimise phasing artefacts, only haplotypes with frequency >5% within each 1 Mb genomic window were retained for subsequent recombination detection.

Given the moderate size of half-sib families, phasing accuracy was a critical consideration. Only recombination events supported by at least two heterozygous informative markers and high posterior confidence were considered for downstream analysis. We acknowledge that small sib sizes limit resolution, but the use of pedigree-based DuoHMM phasing in SHAPEIT2 mitigates this risk by incorporating parental information.

2.5. Detection of Recombination Crossover Events

The DuoHMM algorithm in SHAPEIT2 was used to detect recombination intervals from the phased haplotypes. Noteworthy, haplotypes for each parent–offspring pair were defined by a pair of informative heterozygous markers. Furthermore, recombination rates were estimated for each non-overlapping 1 Mb (assuming 1 cM is equivalent to 1 Mb) window across macrochromosome (1–5), intermediate (6–10), and microchromosomes (11–29) across all the 29 autosomes. The recombination rate in a defined 1 Mb window was computed as

$$C_w=\left({\sum }_{k=1}^n{x}_k/r_k\right)T$$

where Cw was the observed window of recombination rate, and n was the total number of recombination events observed on the corresponding chromosome, x_overlap in Mb between the 1 Mb window and the recombination interval k, r_k the length in Mb of the recombination interval, and T was the total number of parent–offspring pairs. Recombination hotspots were perceived as the SNP intervals with a recombination rate >2.5 standard deviations greater than the mean [18].

2.6. Identification of Genes in Recombination Hotspot Regions

Genomic coordinates corresponding to recombination hotspots were queried against the cattle QTL database (https://www.animalgenome.org/cgi-bin/QTLdb/BT/index, accessed on 12 September 2024) to identify genes located within high-recombination intervals. The gene content of these regions was examined in the context of reported quantitative trait loci (QTL) to infer potential associations with traits relevant to growth, fertility, and environmental adaptability.

3. Results

3.1. Crossover Events

Crossover (CO) events were detected in 1 Mb windows across all 29 autosomes for both Bonsmara and Nguni cattle populations. Notably, Nguni exhibited a higher total number of CO events (n = 11,472) than Bonsmara (n = 6501), despite having fewer genotyped individuals (Figure 1).

In both populations, macrochromosomes (Chr1–Chr5) demonstrated higher crossover counts compared to intermediate (Chr6–Chr19) and microchromosomes (Chr20–Chr29). Chromosome-specific analysis revealed that the highest CO frequency in Nguni occurred on chromosome 1, while in Bonsmara, the peak was observed on chromosome 2.

3.2. Determining Recombination Rates

Breed-specific recombination rates were calculated based on 45,402 markers in Bonsmara and 46,124 in Nguni. Despite the higher number of CO events in Nguni, the average recombination rate was significantly lower (0.18 cM/Mb) than in Bonsmara (0.31 cM/Mb) (Figure 2) (p < 0.01). This paradox may be partly attributed to differences in SNP density and the distribution of informative meioses across chromosomes.

Additionally, the overall recombination rates observed in both populations are substantially lower than the ~1 cM/Mb genome-wide average reported in cattle [17,18,22], suggesting potential underestimation due to marker sparsity or phasing errors.

Table 1. Summary of recombination activities per chromosome across all autosomes.

Chromosome	Bonsmara			Nguni
	#SNP	CO	RR (SD)	#SNP	CO	RR (SD)
1	2369	427	0.29 (0.20)	2806	857	0.18 (0.10)
2	2085	466	0.30 (0.20)	2356	684	0.18 (0.08)
3	1860	292	0.02 (0.24)	2187	617	0.18 (0.009)
4	1751	314	0.32 (0.26)	2095	505	0.18 (0.10)
5	1609	265	0.30 (0.23)	1877	681	0.17 (0.08)
6	2232	349	0.34 (0.22)	2531	687	0.19 (0.11)
7	1948	268	0.33 (0.23)	2230	632	0.19 (0.10)
8	1696	340	0.29 (0.19)	1993	521	0.18 (0.10)
9	1570	216	0.30 (0.24)	1774	489	0.18 (0.10)
10	1768	277	0.31 (0.22)	2063	518	0.19 (0.11)
11	1589	289	0.29 (0.19)	1837	486	0.18 (0.09)
12	1231	209	0.34 (0.25)	1439	434	0.19 (0.09)
13	1232	248	0.33 (0.25)	1470	397	0.18 (0.10)
14	1360	185	0.33 (0.24)	1486	383	0.18 (0.09)
15	1259	230	0.32 (0.22)	1447	363	0.18 (0.11)
16	1192	211	0.31 (0.22)	1418	386	0.18 (0.11)
17	1207	197	0.31 (0.22)	1395	326	0.19 (0.12)
18	962	185	0.30 (0.19)	1116	252	0.18 (0.04)
19	1052	149	0.35 (0.24)	1171	242	0.20 (0.12)
20	1199	175	0.32 (0.24)	1394	371	0.19 (0.11)
21	1073	202	0.31 (0.21)	1234	298	0.17 (0.12)
22	931	156	0.30 (0.20)	1068	243	0.19 (0.19)
23	840	115	0.38 (0.24)	977	184	0.18 (0.20)
24	898	151	0.34 (0.22)	1064	247	0.17 (0.10)
25	743	101	0.26 (0.19)	787	175	0.19 (0.08)
26	748	122	0.25 (0.20)	889	189	0.18 (0.13)
27	723	113	0.32 (0.23)	807	182	0.18 (0.11)
28	674	102	0.30 (0.25)	749	203	0.17 (0.08)
29	761	147	0.30 (0.20)	865	202	0.17 (0.10)

#SNP, number of single nucleotide polymorphism; CO, crossover count; RR, recombination rate.

shows the summary of recombination activities between the two breeds. It can be observed that more CO counts were present in macro chromosomes that consisted of higher SNP density as well. However, on microchromosomes with lower SNP density, the CO counts were not as prominent. Despite this, higher recombination rates were observed on chromosomes with lower SNP density and CO counts.

3.3. Recombination Hotspots Across Autosomes

Recombination hotspot regions were unevenly distributed across the genome and differed markedly between breeds. A total of 407 hotspots were detected in Bonsmara and 179 in Nguni. Interestingly, the chromosome with the highest number of hotspots in Bonsmara was Chr5 (macrochromosome), while in Nguni it was Chr18 (intermediate chromosome) (Figure 3). Across most chromosomes, Nguni harboured more hotspots than Bonsmara, although this trend did not hold uniformly across the genome.

Some chromosomes exhibited unexpected hotspot patterns. For example, Chr9 in Bonsmara, which had relatively few CO events and markers, showed the same number of hotspots as Chr1, which had the highest SNP coverage and CO counts. Similarly, Chr18 in Nguni—despite having fewer markers and CO events—displayed more hotspots than any other chromosome. These findings suggest that SNP distribution and haplotype resolution likely influenced hotspot detection.

3.4. Variation in Recombination Hotspots Along the Autosomes

The distribution of recombination hotspots along autosomes varied by chromosome and breed. In Bonsmara, chromosomes with elevated recombination activity (e.g., Chr1, Chr5, Chr8, Chr9) tended to cluster hotspots in sub-telomeric and telomeric regions (Figure 4). However, Chr1 uniquely exhibited hotspot clustering in both telomeric and centromeric regions.

This suggests potential breed-specific differences in chromatin accessibility or recombination regulatory elements. In Nguni, chromosomes such as Chr1, Chr3, Chr6, and Chr18 showed a broader hotspot landscape, including both telomeric and centromeric enrichments (Figure 5).

3.5. Identification of Genes Associated with Recombination Rates and Hotspots

The functional annotation of genes co-located with recombination hotspots (

Table 2. Analysis of genes in different chromosomes with concentrated recombination hotspots in Bonsmara.

Chr	Position	Gene	Gene ID	Function
1	8.35	AAP	ENSSBTG00000001534	Regulates synopsis formation. Receptor-binding activity. Growth releasing hormone.
1	30.74	ROBO1	ENSBTAG00000009851	Enables protein binding. Enables axon guidance receptor.
1	80.0	BRWD1	ENSBTAG00000025910	Enables molecular function. Regulates transcription by RNA polymerase II. Enables gametogenesis and oocyte embryo transition (Mice). Epigenomic mediator of normal neurodevelopment (mice).
1	75.69	ZBTB38	ENSBTAG00000040061	Regulates cytokine production. Regulates immune system process. Enables protein binding. Negative regulation of transcription by RNA polymerase II. Target gene for body measurement trait.
5	28.76	PDE1B	ENSBTAG00000004337	Target gene for carcass traits.
5	4.60	LOC617565 LINE-1		Determinant for DNA target structure.
8	65.57	Prnp	ENSBTAG00000048903	Regulation of DNA-binding transcription factor. Regulation of long-term synaptic potentiation. ATP independent protein, amyloid and copper binding.
8	103.28	AMBP	ENSBTAG00000015676	Supports ovulation. Repairs micro molecules.
9	91.09	CNKSR3	ENSBTAG00000010581	Regulates sodium ion transport. Maintenance of transepithelial sodium transport in the kidneys in humans.
9	87.84	SCAF8	ENSBTAT00000045257	Enables mRNA processing and RNA transcription. Enables domain specific binding activity.
9	87.84	TIAM2	ENSBTAG00000049782	Encodes Guanine nucleotide exchange factor. Regulates lymphoma. Promotes lung and pancreatic tumour progression in humans.

and

Table 3. Analysis of genes in distinct chromosomes with pronounced recombination hotspots in Nguni.

Chr	Position (Mb)	Gene	Gene ID	Function
1	6.35	PAK2	ENSBTAG00000038996	Chromosomal alignment in metaphase I.
1	25.33	SLC51A	ENSBTAG00000022314	Enables transporter activity of bile acid and bile salt.
1	117.62	BRWD1	ENSBTAG00000023175	Enables molecular function. Regulation of transcription by RNA polymerase II. Enables gametogenesis and oocyte embryo transition in mice. Epigenomic mediator of normal neurodevelopment in mice.
1	62.28	LLP	ENSBTAG00000046064	Responsible for cell adhesion and cell motility.
3	62	FP	ENSBTAG00000018542	Milk fat percentage.
3	38.84	COL241A1	ENSBTAG00000024802	Enables extracellular structure constituent.
3	12.42	CLCA2	ENSBTAG00000035503	Enables intracellular calcium activated chloride channel activity.
6	42.78	NCAPG	ENSBTAG00000025573	Mitotic chromosome condensation.
6	77.34	ADGRL3	ENSBTAG00000032411	Enables G-protein coupled receptor activity.
18	51.42	Prnp	ENSBTAG00000048903	Regulation of DNA-binding transcription factor. Regulation of long-term synaptic potentiation. ATP independent protein, amyloid and copper binding.

) identified several genes involved in meiotic recombination, synaptonemal complex assembly, chromosomal alignment, and reproductive traits such as ovulation and milk fat percentage. A notable finding was the identification of the prion protein (Prnp) gene in both breeds, although it was located on different chromosomes. These observations suggest that conserved and breed-specific genetic elements may drive recombination activity in the South African cattle genome.

Table 2 shows a representation of genes located on specific bovine chromosomes (1, 5, 8, and 9) in Bonsmara cattle that are situated within recombination hotspot regions and are functionally relevant to various biological processes. On chromosome 1, genes such as AAP, ROBO1, BRWD1, and ZBTB38 are involved in transcription regulation, immune processes, gametogenesis, and growth traits. Chromosome 5 includes PDE1B, a gene associated with carcass traits, and LOC617565 LINE-1, which is linked to DNA structural elements. On chromosome 8, Prnp and AMBP are implicated in transcription regulation, ovulation, and molecular repair. Chromosome 9 contains CNKSR3, SCAF8, and TIAM2, genes involved in sodium ion transport, mRNA processing, and cancer-related signalling pathways. These genes highlight the biological significance of recombination hotspot regions and their potential relevance to economically important traits and physiological functions in cattle.

Table 3 also shows a representation of genes located within recombination hotspot regions across several bovine chromosomes (1, 3, 6, and 18) in Nguni cattle, along with their annotated functions. On chromosome 1, PAK2 is involved in chromosomal alignment during metaphase I, SLC51A facilitates bile acid transport, and BRDW1 is associated with transcriptional regulation, gametogenesis, and neurodevelopment (as demonstrated in mice). The LLP contributes to cell adhesion and motility. Chromosome 3 features genes such as FP, which is linked to milk fat percentage, COL24A1, involved in extracellular matrix structure, and CLCA2, which enables calcium-activated chloride channel activity. On chromosome 6, NCAPG plays a role in chromosome condensation during mitosis, while ADGRL3 functions as a G-protein coupled receptor. On chromosome 18, Prnp is again highlighted for its role in transcriptional regulation, synaptic potentiation, and amyloid binding. Collectively, these genes are associated with diverse physiological processes and production traits, reinforcing the functional significance of recombination hotspots in cattle genomes.

In addition to identifying genes within recombination hotspot regions on chromosomes 1, 3, 6, and 18, Table 3 reveals partial overlap with genes listed in Table 2. Notably, the gene BRWD1 is present in both breeds and located on chromosome 1, where it is implicated in transcriptional regulation, gametogenesis, and neurodevelopment. This represents a confirmed gene overlap within the same chromosomal context in both breeds. Additionally, Prnp appears in both breeds; however, it is located on chromosome 8 in Bonsmara and on chromosome 18 in Nguni.

4. Discussion

Breed-specific recombination patterns in beef cattle provide crucial insights into the genomic mechanisms that underpin allelic diversity, phenotypic expression, and breed adaptation. Recombination facilitates the reshuffling of alleles through meiotic crossover (CO) events, promoting the emergence of novel haplotypes that may influence economically important traits such as fertility, carcass quality, disease resistance, and milk production. Therefore, the detailed characterisation of recombination dynamics across genetically divergent breeds not only enhances our understanding of genome evolution but also has direct applications in genomic selection and precision breeding strategies.

This study profiled the genome-wide recombination landscape in two South African cattle breeds, Bonsmara and Nguni, using pedigree-based phased SNP data. While Nguni cattle exhibited a higher total number of CO events (11,472 vs. 6501 in Bonsmara), the average recombination rate was significantly higher in Bonsmara (0.31 cM/Mb) than in Nguni (0.18 cM/Mb). This inverse relationship suggests that although Nguni cattle undergo more recombination events, these events are more diffusely distributed across the genome, resulting in lower recombination intensity per Megabase. Such disparity may reflect breed-specific differences in chromatin accessibility, sequence motif distribution, or the regulatory influence of recombination modifiers such as PRDM9. The concentration of COs in macro- and intermediate-sized chromosomes in both breeds confirms the established correlation between physical chromosome length and recombination frequency [17,27]. Notably, chromosomes 1 and 2 harboured the highest crossover densities in both breeds, consistent with observations in taurine breeds such as Holstein. This size-dependent crossover distribution aligns with the “obligate crossover” hypothesis, which posits that larger chromosomes require more COs to ensure proper meiotic segregation [9].

Despite these shared recombination patterns, pronounced breed-specific differences emerged in the number and distribution of recombination hotspots. Bonsmara harboured 407 hotspots, more than twice the number observed in Nguni (179). These findings underscore that while total CO counts reflect genome-wide recombination processes, hotspot density is shaped by localised genomic features and may serve as a proxy for recombination efficiency. In Bonsmara, hotspots were predominantly concentrated in telomeric regions, although centromeric enrichment was also detected, particularly on chromosome 1. In contrast, Nguni cattle exhibited a more dispersed hotspot landscape with both centromeric and telomeric enrichment. This pattern is rarely reported in taurine breeds and may indicate unique epigenomic architectures in Sanga-type cattle. The presence of centromeric recombination hotspots in Nguni contrasts with canonical mammalian patterns, which typically suppress recombination in centromeric regions due to their heterochromatic nature. This deviation may reflect breed-specific epigenetic modifications or the presence of recombination-promoting motifs in these regions. Alternatively, it may result from historical admixture, environmental selection pressures, or structural variations in centromeric DNA sequences that facilitate recombination events in otherwise recombination-suppressed regions. These warrant further investigation using functional genomic assays and chromatin accessibility mapping in indigenous African breeds.

The observed hotspot counts align with published data for other breeds: the total in Bonsmara is comparable to Angus (424) and Limousin (348) cattle [21], while Nguni exhibits a comparatively lower density. This difference may be biologically relevant, given the Nguni breed’s evolutionary history as indigenous Sanga-type cattle adapted to diverse and harsh environmental conditions. However, it could also be partially artefactual, due to the smaller sample size and family structures in the Nguni dataset, which reduce the statistical power to detect true COs and increase the risk of phasing errors. Therefore, while comparisons of recombination landscapes across breeds are biologically informative, methodological limitations, particularly family size, marker density, and SNP resolution, must be carefully considered when interpreting breed-specific findings.

The gene content of hotspot-enriched regions provides compelling evidence for functional relevance, yet caution is warranted when inferring causality between gene function and recombination activity. Several genes associated with meiosis, chromosomal alignment, and reproductive function were detected in both breeds. Among these, Prnp was identified within hotspot regions in both Bonsmara (Chr8) and Nguni (Chr18), although with differing chromosomal localisation. This discrepancy likely reflects differences in genome annotation or structural variation between the reference genomes but supports a conserved role for Prnp in reproductive processes and potentially in recombination regulation. Another shared gene, BRWD1, was detected in both breeds on chromosome 1 and is known to facilitate gametogenesis and oocyte–embryo transition, further corroborating its potential involvement in meiotic recombination. Additionally, the identification of SCAF8 in Bonsmara, a gene known to regulate PRDM9, the master controller of recombination hotspot localisation, suggests that PRDM9-mediated recombination may be more active or differently regulated in this breed. Given that PRDM9 binding motifs determine hotspot specificity, differential expression or sequence variation in SCAF8 or PRDM9 could partially explain the higher hotspot density observed in Bonsmara. However, functional validation through gene expression profiling, PRDM9 motif analysis, and epigenetic mapping is required to substantiate this hypothesis.

Several genes located within hotspot regions in Bonsmara and Nguni are associated with economically important traits, providing a potential link between recombination dynamics and phenotypic selection. For example, PDE1B, implicated in carcass traits, and FP, linked to milk fat percentage, were identified within recombination hotspots. The co-localisation of recombination hotspots with regions harbouring trait-associated genes raises the possibility that recombination could facilitate the generation of favourable haplotype combinations under selection. To strengthen this interpretation, future studies should integrate recombination hotspot maps with quantitative trait loci (QTL) databases, such as the Animal QTLdb, to assess overlap between hotspots and regions influencing fertility, growth, meat quality, disease resistance, and climate adaptability. Such integrative analyses would enable a more robust exploration of the causal relationships between recombination patterns and trait evolution, moving beyond descriptive hotspot mapping towards functional genomics and trait-association frameworks.

Furthermore, the atypical localisation of recombination hotspots in Nguni, particularly within centromeric and pericentromeric regions, may reflect signatures of adaptation or structural genome differences unique to this indigenous breed. This pattern could represent an evolved mechanism to maintain genomic integrity or facilitate recombination under environmental stress, supporting the hypothesis that indigenous African cattle possess distinct genomic architectures shaped by long-term natural and artificial selection in resource-limited and variable climates. Recombination in centromeric regions may also contribute to enhanced genetic diversity in Nguni populations, potentially buffering against inbreeding depression and promoting resilience in low-input production systems. These findings underscore the need for comprehensive structural variation and epigenetic profiling in Nguni and other Sanga-type breeds to elucidate the underlying mechanisms of recombination regulation.

From an applied perspective, the integration of recombination architecture into breeding programmes offers tangible opportunities to enhance genomic selection models in South African beef cattle. The higher recombination intensity observed in Bonsmara may lead to smaller haplotype blocks, thereby improving the resolution of genomic predictions and facilitating the fine-mapping of causal variants. Conversely, the distinct recombination landscape of Nguni, characterised by fewer but more dispersed hotspots and centromeric activity, highlights the necessity of developing breed-specific genomic tools that account for unique linkage disequilibrium (LD) structures. This is particularly relevant for crossbred or indigenous populations where reliance on taurine-derived reference panels may compromise predictive accuracy. By incorporating recombination metrics into genomic evaluation models, breeding programmes can optimise marker density selection, improve genomic estimated breeding values (GEBVs), and refine selection indices to balance productivity and adaptation traits.

Overall, this study represents the first genome-wide characterisation of recombination in South African Bonsmara and Nguni cattle. It reveals distinct patterns in recombination rates, crossover frequency, and hotspot localisation between the two breeds, providing novel insights into the evolutionary and functional genomics of African beef cattle. Future research should leverage high-density genotyping arrays or whole-genome sequencing, combined with larger and more balanced pedigree structures, to validate these findings and unlock the full potential of recombination-driven genomic improvement. Moreover, integrative approaches that combine recombination maps, QTL data, gene expression profiles, and chromatin state analyses will be essential to elucidate the biological significance of recombination hotspots and translate this knowledge into actionable strategies for sustainable genetic improvement in African livestock populations.

5. Conclusions

This study provides insights into the recombination landscape of two South African beef cattle breeds, Bonsmara and Nguni, using pedigree-based phased genotype data. The findings suggest the presence of breed-specific patterns in recombination rate and hotspot distribution, with notable differences observed between the composite taurine Bonsmara and the indigenous Sanga-type Nguni. These results add to the growing body of evidence that recombination is not uniformly conserved across cattle breeds, underscoring the potential for genomic diversity in African livestock to contribute to broader recombination research. A key contribution of this study lies in the characterisation of recombination architecture in an underrepresented group of African breeds, particularly the Nguni, for which such genomic analyses are limited. The detection of recombination hotspots distributed within centromeric as well as telomeric regions in Nguni cattle represents a potentially novel recombination signature that may warrant further investigation. Additionally, the identification of meiosis-related genes located in recombination-active regions supports the biological plausibility of the observed patterns.

However, these findings must be interpreted with thoughtfulness given the inherent limitations of the dataset. The study relied on half-sib family structures with relatively small family sizes, which may reduce the resolution and accuracy of recombination rate and hotspot inference. Furthermore, the use of SNP array genotype data, while informative, may not capture the full extent of recombination activity compared to whole-genome sequencing or denser genotyping platforms. The limited number of informative meioses, especially in the Nguni population, may have resulted in underestimation or mischaracterisation of true recombination events. Future studies should prioritise the generation and analysis of more robust datasets, ideally based on full-sib or trio-based family structures and incorporating larger sample sizes. Such datasets would facilitate higher-resolution mapping of recombination events and enable a more accurate assessment of recombination landscape heterogeneity across breeds. The integration of whole-genome sequence data and functional annotation of recombination-associated loci would also enhance the interpretability of recombination hotspots and their potential linkage to traits of economic importance.

While this study contributes new information to the understanding of recombination in African beef cattle, it also highlights the need for continued investigation using more powerful and representative data. Future work should aim to refine these observations and explore their relevance for genomic selection and breeding strategies in African livestock systems.