Genetic characterisation of non-descript cattle populations in communal areas of South Africa

Additional keywords: genetic diversity, heterozygosity, microsatellite markers.

Indigenous breeds are important genetic resources for the livelihood of rural societies (Ramsay et al. 2000) to meet nutritional, economic and socio-cultural requirements (Zulu 2009). These indigenous breeds have an ability to withstand harsh climatic conditions (Thornton et al. 2009), resistance to parasites (Mapholi et al. 2016), diseases (Marufu et al. 2011, 2014) and genetic adaptation to poor-quality forages. The unique gene pool is under threat because of alterations in farming systems (Kristensen et al. 2015; Marsoner et al. 2017), such as crossbreeding occurring in small population sizes, the absence of breed societies (whose role is to keep records of pedigrees, promote and encourage the conservation, breeding and genetic improvement of the production potential among other things) and inbreeding depression (Ollivier and Foulley 2002; FAO 2007). These factors can lead to possible extinction or a total loss of a breed (Ollivier and Foulley 2002); therefore, it is important that the farm-animal genetic resources are managed effectively for future use.

Genetic-diversity studies conducted in South Africa have focused on stud and commercial herds of the Afrikaner cattle (Pienaar et al. 2014, 2018), six cattle breeds from research stations (Makina et al. 2014) and Nguni cattle ecotype (Sanarana et al. 2016). The genetic status of cattle populations existing in communal areas of South Africa is unknown and implementation of appropriate conservation measures must be considered to ensure the effective management of indigenous-animal genetic resources (Taberlet et al. 2008; Boettcher et al. 2010; Makina et al. 2014). With the constant decline in farm-animal genetic resources (Rege 1999; Nyamushamba et al. 2017; Gwaze et al. 2009), there is a need to characterise beef cattle populations of the communal area, so as to understand the existing diversity to facilitate the development of rational conservation strategies and sustainable utilisation for the breeds (Hanotte and Jianlin 2006). Genetic characterisation of these populations will provide a better understanding of breed formation that can be used in selection program for potential animals for present and future genetic resources (Sunnucks 2000; Toro and Caballero 2005; Tixier-Boichard 2014).

Traditionally, characterisation of the non-descript populations was based on their phenotypic information such as, for example, body frame and coat colour (Teneva 2009). However, as a result of the recent advancement in genomics technology, characterisation is now conducted using genetic information to improve their production. In recent years, the use of genetic markers such as microsatellites to determine the genetic diversity of populations, has gained popularity (Vignal et al. 2002; DeSalle and Amato 2004), including mapping of genes controlling economically important traits has been made possible by the use of microsatellite markers (Beuzen et al. 2000). With the use of genetic markers, it is possible to test for parentage, relatedness of a population, an individual’s identification, as well as to determine whether any migration has occurred (Hanotte and Jianlin 2006; Toro et al. 2009). Microsatellite markers have been used to characterise the diversity of South African chickens (Van Marle-Köster et al. 2008; Mtileni et al. 2011), cattle (Pienaar et al. 2014; Sanarana et al. 2016), sheep (Soma et al. 2012; Qwabe and Van Marle-Köster 2013) and goat (Mdladla et al. 2017). Therefore, the study was conducted to assess the genetic diversity and population structure of South African non-descript communal beef cattle populations, by using 25 microsatellite markers.

In total, 150 hair samples were randomly collected from different farmers from five non-descript cattle populations. This was from the Eastern Cape (EC, n = 30), KwaZulu–Natal (KZN, n = 30), Limpopo (LP, n = 30), Mpumalanga (MP, n = 30) and the North West (NW, n = 30) provinces of South Africa. Since there is no well structured recording system in smallholder farms, unrelated female animals (one per farmer) were selected and sampled. The sampled hair was collected from the tail end of each animal. Each sample was placed in an individual bag, sealed and labelled as per animal identification, so as to avoid any contamination. Six known beef cattle breeds were obtained from the Agricultural Research Council database, and used as reference populations for genetic admixture evaluation. The reference populations included Angus (ANG), Afrikaner (AFR), Bonsmara (BON), Brahman (BRA), Drakensberger (DRA) and the Nguni (NGU), with these being 30 animals per breed. The study was conducted in accordance with the approval of the Ethics Committee (AEC) of the Agricultural Research Council, South Africa (APIEC17/08).

DNA was extracted from each hair sample with visible roots, by using phenol–chloroform, following the protocol of Sambrook et al. (1989). The concentration (260/280 nm) and the purity (260/230 nm) of the genomic DNA was measured using a spectrophotometer (Nanodrop 2000c; Thermo Fisher Scientific Inc., Waltham, MA, USA; and Nanodrop 2000c). DNA polymorphisms were determined using a set of 25 autosomal microsatellite markers (Table 1) recommended by the Food and Agricultural Organisation of the United Nations and the International Society for Animal Genetics Advisory Group (FAO 2011) for genetic-diversity studies. These markers were selected on the basis of their high level of polymorphism reported by several studies (Kim et al. 2004; Pienaar et al. 2014; Sanarana et al. 2016). The DNA amplification was performed using a Perkin Elmer Gene Amp PCR System 9700 Thermo cycler (Applied Biosystems, Foster city, CA, USA). Amplicons were separated by capillary electrophoresis using an ABI 3130xl automatic sequencer (Applied Biosystems). Fluorescently labelled fragments were detected and sized using GeneMapper software (version 4.0, Applied Biosystems).

Table 1.  Allele-size ranges, number and chromosome location of alleles per population
EC, Eastern Cape province; KZN, KwaZulu–Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province

The genetic diversity per locus and across the populations was estimated using Microsatellite Toolkit (Park 2001). Arlequin version 3.1 was used to perform locus by locus analysis of molecular variance (AMOVA), so as to determine the differentiation within and among the populations, genetic measures per locus and population (Excoffier et al. 2005). Genetic relationship among communal beef cattle populations were determined according to Nei’s standards (Nei 1987).

The genetic population-structure analysis of the communal cattle was performed using Bayesian admixture procedure implemented in STRUCTURE 2.3.4 (Pritchard et al. 2000), to infer the most likely number of clusters. The most probable number of populations was determined according to Evanno et al. (2005). A length of burning period was set at 10 000, with the number of Markov-chain Monte Carlo (MCMC) reps after burning being 20 000 and the number of iterations being 14 (for 2 ≤ number of clusters (K) ≥ 20). The most probable K value that reasonably describes the substructure of the populations under study was determined from the log probability of the data (Ln Pr (X|K)), using the STRUCTURE Harvester software (Earl and Von Holdt 2012), which implements Evanno’s method (Evanno et al. 2005). All other parameters in STRUCTURE were left as default (the default initial value of α was 1.0, so the value of λ was 1.0). The clustering pattern was visualised using DISTRUCT 1.1 (Rosenburg 2004). The factorial correspondence analysis (FCA) was computed using DARwin ver. 6 software package (Perrier and Jacquemound-Collet 2006). POPTREE2 was used to construct the phylogeny of the populations (Takezaki et al. 2010).

Of 25 microsatellite markers used to study the genetic diversity of five non-descript cattle populations from five provinces of South Africa, 24 were found to be polymorphic, except one marker (HAUT27). Table 2 shows measures of genetic diversity among and within the populations. The average number of alleles ranged from 7.68 ± 2.10 for MP to 8.48 ± 2.40 for KZN. The expected heterozygosity (H_E) ranged from 0.72 ± 0.03 (EC) to 0.79 ± 0.02 (LP), while observed heterozygosity (H_O) ranged from 0.56 ± 0.02 (EC) to 0.70 ± 0.02 (LP and NW). All H_E values were higher than the H_O values in all populations. Moderate levels of inbreeding were observed in all populations, ranging from 0.02 to 0.20.

Table 2.  Genetic diversity measures of the five populations
F_IS, within-population inbreeding estimates; H_E, unbiased heterozygosity; H_O, observed heterozygosity; EC, Eastern Cape province; KZN, KwaZulu–Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province

The AMOVA showed 10% of variation among populations and 11% of the variation among individuals, with the remaining 79% accounting for the differences within individuals in the population (Table 3).

Table 3.  Analysis of molecular variance from the five populations

Nei’s genetic distances among populations are shown in Table 4. The highest genetic differentiation of 0.42 was observed between MP and EC, while the lowest genetic differentiation of 0.02 was between EC and KZN.

Table 4.  Pair-wise matrix of Nei’s genetic distance of the five populations
EC, Eastern Cape province; KZN, KwaZulu–Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province

Genetic clustering based on STRUCTURE analysis of the five populations is shown in Fig. 1. This was performed following Rosenberg et al. (2004), by using a Bayesian approach that inferred the K present in the population, permitting the detection of differences among populations and the hidden substructure within them. Each population was represented by one column, each with a different colour. The most probable cluster was found at K = 7. The non-descript populations formed three clusters, namely, the first cluster (NW), the second cluster (MP) and the third cluster (EC, KZN and LP). The clustering of EC, KZN and LP populations suggests that these populations are very similar; it is also noted that they had genetic material similar to those possessed by the NGU and BON reference populations. The DRA, AFR, ANG and the BRA reference populations each formed separate clusters.

Fig. 1.  Structure clustering of the five populations assessed against six reference populations at K = 7, where K is the number of clusters present in the population. EC, Eastern Cape province; KZN, KwaZulu-Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province; ANG, Angus; AFR, Afrikaner; BON, Bonsmara; BRA, Brahman; DRA, Drakensberger; NGU, Nguni.

The FCA was performed to determine the genetic relationship among the studied populations and the results are presented in Fig. 2. FCA showed very clear separation between the NW and MP. This analysis also showed an overlap of the EC, KZN and LP populations, suggesting a genetic relationship among the three populations.

Fig. 2.  Factorial correspondence analysis of the cattle populations in the five populations. EC, Eastern Cape province; KZN, KwaZulu-Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province.

The rooted neighbour-joining dendogram illustrated the genetic divergence between the five non-descript populations and the six reference populations (Fig. 3). The EC, KZN, LP shared some genotypes with the reference populations (NGU, DRA, BON, AFR and ANG) as they clustered together. The NW, MP and BRA populations formed two separate clusters. The BRA was the only reference population that formed a separate cluster.

Fig. 3.  Rooted neighbour-joining dendogram, showing the genetic relationships among the studied populations. EC, Eastern Cape province; KZN, KwaZulu-Natal province; LP, Limpopo province; MP, Mpumalanga province; NW, North West province; ANG, Angus; AFR, Afrikaner; BON, Bonsmara; BRA, Brahman; DRA, Drakensberger; NGU, Nguni.

Genetic characterisation is an important tool in the implementation of appropriate breeding programs. It is important to genetically characterise farm animals so as to conserve the existing genetic material (Arora et al. 2008). In the current study, the focus was on genetic characterisation of non-descript-type cattle populations in communal farming in South Africa. This was achieved by assessing the level of genetic variation and inbreeding in the communal beef cattle populations by using autosomal microsatellite markers and by determining the population structure of beef cattle population, by using a reference dataset consisting of other South African purebred commercial lines.

The genetic diversity within the populations is indicated by the average number of alleles observed over a range of loci in different populations (Machugh et al. 1997; Hassen et al. 2012). Average number of alleles is, therefore, a strong indicator for the evolutionary potential of a population (Allendorf et al. 2012; Caballero and Garcıa-Dorado 2013), and it has been suggested that this measure is of key importance in population management and conservation (Simianer 2005; Foulley and Ollivier 2006). The overall average number of alleles in the populations of 8.13 observed in the current study was lower than that obtained in South African Nguni cattle (9; Sanarana et al. 2016) and that in Ankole longhorn cattle in the African Great Lakes Region (13.8; Ndumu et al. 2008). Moreover, the average number of alleles in the present study was higher than that observed in Afrikaner cattle in South Africa (4.8; Pienaar et al. 2018) and that in Yellow cattle in Taiwan (3.8) (Tu et al. 2014). The high average number of alleles in KZN and LP (8.48) is an indication of a higher genetic variation, and the low average number of alleles in MP (7.68) is an indication of a reduction in the population’s potential to adapt to future environmental changes, since this diversity is the raw material for evolution by natural selection (Fisher 1930). Barker (1994) suggested that the required minimum number of alleles for a microsatellite study is four and, thus, this study complied with that number as a higher number of alleles were observed, therefore justifying the selected markers from the FAO panel of microsatellite markers.

Nei (1987) and Barker (2001) reported that gene diversity (average H_E) is a great measure of genetic variation within a population. High genetic variation occurs when there is a high number of alleles and this can be achieved by having a large population size. Genetic diversity ranges between 0 and 1 and this means that the historic admixture of different populations and a long-term natural selection for adaptation are indicated by high gene-diversity levels (Ojango et al. 2011; Paiva et al. 2011). The H_E mean value of 0.75 observed in the current study was lower than the 0.92 observed in Vechur cattle in India (Radhika et al. 2018). The H_O values ranged between 0.56 and 0.70, with a mean value of 0.66, and this was higher than 0.52 observed in Lidia bovine breed in Spain (Canon et al. 2008), but lower than the value of 0.67 for Nguni cattle populations in Mozambique (Bessa et al. 2009) and the value of 0.69 for Kankrej Cattle in India (Sodhi et al. 2007). Therefore, the difference between the H_O and H_E could be due to non-random mating among the individuals of the population (Sharma et al. 2016). The H_E values in the present study were higher than that observed across all populations, and this may be attributed to forces such as inbreeding (Mburu and Hanotte 2005), isolation or genetic drift resulting in loss of genetic diversity (Lacy 1987). The LP population was found to be the most diverse population in the study, with the highest H_E (79%) and H_O (70%), and with an average number of alleles of 8.48.

It was observed that there was a higher genetic variation within individuals in the population (79%) than among populations (10%) or among individuals (11%). These low levels of differentiation among populations suggested common historical origins and a high level of inter-population gene flow (Radhika et al. 2018). These findings are in line with Groeneveld et al. (2010), who stated that most of the genetic diversity is present within a breed and not among breeds. For a breed to adapt to different environments and to respond to selection, genetic diversity is a requirement (Frankham et al. 2002) and AMOVA describes the genetic differentiation within and among populations (Toro and Caballero 2005). AMOVA showed that there was more genetic differentiation that occurred within populations than among populations and this might be due to the fact that there is no selection for or against any specific production trait expressed in the populations.

Nei’s genetic-distance pairwise matrix estimates were used to estimate the genetic relationship among the studied populations and they indicated that the EC and KZN were much closer to each other than they were to other populations. From these results, it can be concluded that farmers in those two provinces could be buying cattle from one another as these provinces are neighbours and transport costs can be kept at a minimum.

The results of STRUCTURE, FCA as well as the neighbour-dendogram analyses were in line with each other as they all supported the genetic distinctiveness of the NW and the MP populations. The NW and MP populations distinctively clustered away from each other and from any of the studied populations in all the methods that were used. The population STRUCTURE allows the evaluation of a population’s admixture from individuals, migrants in a population and also evaluates whether crossbreeding has occurred (Pritchard et al. 2000). The EC, KZN and LP populations had genetic material similar to that of the NGU and BON reference populations. Reasoning behind the EC, KZN and LP sharing genetic material similar to that of the NGU and BON is because these populations (EC, KZN and LP) are being reared in the communal areas, where there is no proper breeding programs, wherein any dominant male has a chance to breed (Scholtz et al. 2008). Most cattle in the rural areas are non-descript crossbreds with small populations of local breeds, such as Afrikaner, and locally developed Bonsmara and Drakensberger (Palmer and Ainslie 2006). The NGU and BON populations clustering together could be due to the fact that they are admixed with a distinct genome component that is shared with the Kuri (a hybrid between indicine populations and African taurine) and Ankole–Watusi, Boran, Sheko, short-horned Zebu and zebu Bororo (African zebus; Gautier et al. 2009). This was expected as most of the cattle in the country, although not being genetically verified, are regarded as NGU, BON or AFR and this could be due to the fact that most communal farmers practice uncontrolled mating programs and the animals graze together with those of NGU, BON or AFR farmers in the regions. Steinfeld et al. (2006) stated that strong management systems and agricultural practices are required for the conservation of livestock biodiversity. Another reason could be that farmers often prefer farming with the Nguni cattle, mainly because of the characteristics they possess (Baker and Rege 1994; Nyamushamba et al. 2017; Shabtay 2015). Correspondingly, the factorial analysis method showed a close genetic relationship among the EC, KZN and LP populations as these populations were mixed but the NW and MP populations formed separate clusters respectively. It is evident that the non-descript populations have an admixture of genetic material in common with those of the reference populations (the NW and MP excluded). The EC, KZN and LP populations had genetic material similar to that of any reference population, namely, the NGU and BON populations. This suggested that the EC, KZN and LP populations possessed similar genetic material.

The present study was the first to attempt to evaluate the existing genetic diversity of non-descript cattle populations in the communal areas of South Africa. The studied populations that were found to have genetic material similar to that of any of the reference populations were the EC, KZN and LP. The NW and MP populations did not possess any genetic material that was similar to that of the reference populations. Correct breeding programs should be put in place to conserve the current genetic material. The genetic diversity in the studied areas can be used as a baseline in the conservation, development and implementation of breeding programs. Programs that are directed to control indiscriminate crossbreeding practices in smallholder systems because an admixture of animals has been shown to suppress production, due to the mismatch among the production, environment and genotype, so as to lower the possible genetic dilution of the current available genetic material.

The authors declare no conflicts of interest.