Genetic diversity of local cattle

1Slovak University of Agriculture in Nitra, Faculty of Agrobiology and Food Resources, Department of Animal Genetics and Breeding Biology, Nitra, Slovak Republic; 2University of Zagreb, Department of Animal Science, Zagreb, Croatia; 3University of Ljubljana, Department of Animal Science, Ljubljana, Slovenia; 4University of Natural Resources and Life Sciences, Division of Livestock Sciences, Vienna, Austria


INTRODUCTION
Threats to biodiversity in terms of extinction rate, destruction of ecosystems and habitat, or loss of genetic diversity are increasing within livestock.The current concept of breed was formulated in the mid-1800s, followed up by intense artificial selection leading to striking and large-scale changes in the genetic makeup of individuals and the production systems.In this process, the farmers progressively substituted the less productive, lo-cally adapted, native breeds with highly productive cosmopolitan breeds, abandoning marginal production areas.As a result of this process, a significant number of cattle, sheep, and goat breeds has already disappeared and many became endangered (Taberlet et al., 2008).
The genetic structure of natural populations is influenced by a limited gene flow occurring when geographical distances increase between them.Non-random mating is a result of the preferential mating between geographically close individuals; this means that the genetic distance between individuals or populations is proportional to their spatial distance ("isolation by distance").The isolation by distance models are used to study demographic, migratory reproductive aspects of populations (Loiseau et al., 2009).Gene flow in natural populations depends on ecological characteristics and global or local environment, while in the case of livestock it is mainly influenced by human activities (Allendorf et al., 2010).
In Central Europe, most breeds were of triple purpose.In remote Alpine valleys, productivity was improved only in the late 19 th century by better feeding and management.Red-pied breeds such as the Dutch Meuse-Rhine-Yssel were more developed as dual purpose cattle, but in the late 20 th century were influenced by red-pied Holstein sires.Successful breeds had spread beyond their region of origin and were even exported to other countries or continents.In contrast, locally adapted but less productive breeds declined in number or disappeared.Genetic isolation of breeds decreased the diversity at the molecular level, which can be monitored via an increase in homozygosity (Felius et al., 2014).
The Slovak Pinzgau breed is endangered cattle in which selection with intense and highly unbalanced use of superior bulls magnified the loss of genetic diversity.The genetic gain increases along with the inbreeding coefficient, which may cause problems due to its small population size.Change of breeding goal to preserve the dual purpose character of Slovak Pinzgau is proposed, with an expected positive impact on the population structure (Kasarda et al., 2014).Polymorphism of genes encoding leptin, leptin receptor (Trakovická et al., 2013) and five others candidate genes associated with production parameters (Miluchová et al., 2014) was evaluated in the Slovak Pinzgau cattle, with the determination of genotypic population structure using molecular genetic methods.The use of most recent results of quantitative and molecular genetics should help to sustainably manage population.With application of genomic selection, a routine collection of genomic information provides an invaluable tool for effective management of breeding programs in small, endangered populations (Mészáros et al., 2015).
The main objective of this study was to provide a detailed insight into genomic characterisation of the Pinzgau cattle (Slovak and Austrian Pinzgau) by utilizing high-throughput molecular information.Bayesian clustering algorithms were used to detect gene flow among the Pinzgau breeds, Cika and Piedmontese cattle.

MATERIALS AND METHODS
The analysed dataset consists of 179 cattle genotypes (19 Slovak Pinzgau, 105 Austrian Pinzgau, 26 Cika, and 29 Piedmontese).Slovak and Austrian Pinzgau cattle are dual purpose Alpine cattle breeds suitable for mountain regions, with good longevity, fertility and adaptability to poor production conditions.Recently, the Slovak Pinzgau cattle population size has declined due to the crossing and replacement with highly intensive breeds, e.g.Red Holstein-Friesian cattle.
The data sets were merged and quality controlled, SNPs with average heterozygosity >0.5, with call rate <0.8, with significant deviation from Hardy-Weinberg equilibrium at P<10 -5 were removed, following McTavish and coworkers (McTavish et al., 2013a).All individuals with >10% missing SNPs placed at 29 autosomal chromosomes and all SNPs that were missing in >10% of individuals were removed.Minor allele frequency threshold of 0.05 was imposed in all calculations.The LD-based pruning was performed to remove SNPs that were in linkage disequilibrium (LD) with one another by using PLINK's (Purcell et al., 2007) simple pairwise threshold model (command: --indep-pairwise 50 5 0.05).This removed 25,433 SNPs that exceeded LD threshold of 0.05, after which 9,433 SNPs remained for further analyses.
Bayesian approach using BAPS version 6.0 (Corander et al., 2008) was used to find admixture and to produce gene flow network.The optimal number of clusters was directly inferred by the implemented algorithm.The maximum number of clusters was set to 5, as recommended in case of higher than expected number of populations.An admixture analysis conditional on the optimal genetic mixture estimate from the individual level analysis was performed.Results were based on 1000 simulations from the posterior allele frequencies.The number of clusters containing more than 19 individuals as a point estimate of K was used.Furthermore, to assess the significance of the admixture estimates, 200 individuals were generated from each identified ancestral source (initial breed) to provide an approximation to the distribution of the estimates under the hypothesis of no admixture.Number of iterations for the reference individuals was set to 10.A gene flow network for the clusters estimated from admixture results was produced (Tang et al., 2009).

RESULTS AND DISCUSSION
As a starting point of this study, previous analyses, partly mentioned by Kukučková and coworkers (Kukučková et al., 2017), were used.The degree of the genetic differentiation between analysed cattle populations (SP, AP) was initially investigated.The level of F ST indicating differences between SP and AP was 0.01.The authors had found a low level of genetic differentiation between CK and PI as well (0.04), thus a more thorough study was needed.The similarity of 4 evaluated populations is shown in more detail in Fig. 1.Parameter estimation in metapopulations is more valuable and more accurate (than in isolated populations with no gene flow), and facilitate investigations of the relative importance as well as interactions among drift, selection and migration in populations of conservation concern (Allendorf et al., 2010).
Figure 1B presents 15 exact clusters for the metapopulations of 15 European breeds (K=15) using unsupervised clustering without prior information about populations, based on 1000 simulations from the posterior allele frequencies.On the contrary, Fig. 1A visualised results based on a smaller number of simulations, with the resulting number of clusters K=13.The Pinzgau populations are in one cluster (coloured yellow), while the CK and PI are grouped in another cluster, coloured dark blue.The influence of the Pinzgau cattle in formation of the CK population is also apparent.
According to Felius and coworkers (Felius et al., 2014), the Piedmontese beef breed belongs to the Central European cattle but does not belong to a breed cluster of the spotted dairy cattle from the eastern Alps (Cika and Pinzgau).By using a Bayesian approach, McTavish and coworkers (McTavish et al., 2013) have confirmed the hypothesis that the New World cattle (as well as modern breeds from Southern Europe, e.g.Piedmontese cattle) are influenced by an ancestral gene flow from northern Africa.
To determine the degree of genetic differentiation among the four studied local populations, the SNP chip data were analysed here using a genetic mixture model for unlinked markers.Using BAPS, changes in the log marginal likelihood based on posterior mode estimates of the admixture coefficients (set as significant at the 5% level) are shown in Fig. 2.There are 4 estimated populations (clusters) and cluster 3 was chosen as the 'source'.According to Tang and coworkers (Tang et al., 2009), changes in the log marginal likelihood close to zero indicate that the mixture model judges both assignments (source cluster and target cluster) to be statistically reasonable for an individual.In contrast, values further away from zero indicate that the relative genetic affinity of the clusters has decreased.In Fig. 2, it is apparent that AP is genetically much closer to SP, than CK and PI.The density curve for AP is shifted more towards zero than the other populations, hence it implies a closer relationship to SP.The low estimate of genetic differentiation between SP and AP indicated that differentiation among these populations is low, particularly due to a common historical origin and high gene flow.Analyses of Allendorf (Allendorf, 2010) assume that genetic distance provides an estimation of genetic connectivity, by reflecting the degree to which gene flow and drift affect genetic divergence among populations.
The assignment of individuals from high-throughput genotyping data was achieved using Bayesian inference, confirming that it is possible to distinguish among the breeds by using SNPs and that the breeds are genetically distinct (Fig. 3 Top).To investigate the ancestral admixture of a certain population, one can look at the arrows pointing at this population.A typical population contains major sources of its own gene pool, denoted as a selflooping arrow, and small proportions of gene flow from other populations (Tang et al., 2009).The actual rates of admixture represented by the arrows and corresponding numbers were randomly generated from a uniform distribution (Fig. 3 Bottom).Two ways of gene flow between SP and CK, SP and AP, CK and AP were found.The PI population contributed to each of the analysed populations.This was rather surprising as all populations except PI indicated two ways of gene flow, which means that PI was involved in formation of the analysed breeds, while these breeds were not involved in creation of PI.In numbers, 87% of SP is unique for the breed, while 12% of its genetic information comes from AP, 0.58% from Cika and 0.26% from Piedmontese cattle.
Gene flow in our study is presented as a proportion of genetic information identical for a pair of breeds, while Achmann and coworkers (Achmann et al., 2004) estimated gene flow as a number of migrant horses per generation from a subpopulation and to a subpopulation.Furthermore, the pool of genotypes for the entire breed enables to continuously scan the population and allow a swift reaction in identifying carriers of lethal or potentially harmful haplotypes.The new information can be used to eliminate undesirable alleles through the mating process.Similarly, the breed proportions due to admixture could be estimated with the goal to fix a desirable ratio or to preserve the purity of the breed (Mészáros et al., 2015).
Although the diversity of the current cosmopolitan cattle is still large enough to belie the claim that cattle becomes an endangered species, the loss of local breeds does erode genetic resources that are difficult to replace.Furthermore, these breeds belong to our cultural heritage and are of local cultural importance, even if most of the breeds are only one or two centuries old (Felius et al., 2014).
Nielsen and coworkers (Nielsen et al., 2009) found that adaptive population divergence may be prevalent despite seemingly high levels of gene flow, as found in most marine fishes.Their results have important implications for understanding of the interplay of evolutionary forces in general, and for the conservation of marine biodiversity under rapidly increasing evolutionary pressure from climate and fisheries induced changes in the local environments.

CONCLUSION
We demonstrated the power of using molecular markers obtained by high-throughput genotyping in clustering cattle breeds, even in case of a very weak differentiation, as between AP and SP (F ST =0.012).Changes in the log marginal likelihood also indicated AP as the most similar breed to SP.All populations except the Piedmontese breed, showed two ways of gene flow among populations, which means that PI was involved in formation of the analysed breeds while these breeds were not involved in creation of PI.The obtained genetic data shed light on some issues related to the local Pinzgau breeds' origin and structure.Our study has shown that the Slovak, as well as the Austrian cattle, are important and viable targets for conservation because they display special traits of phenotypic, cultural and historical na-   ture.It is more strategically important than ever to preserve as much of the livestock diversity as possible, to ensure a prompt and proper response to the needs of future generations.In this context, approaches based on the combination of genomics and spatial analyses are of great help.

Figure 3 .
Figure 3. Admixture analysis in the population of Slovak Pinzgau (SP -Cluster 1), Piedmontese (PI -Cluster 2), Cika (CK -Cluster 3) and Austrian Pinzgau (AP -Cluster 4).(Top) Each column represents a single individual and is coloured according to the proportion of genetic information assigned to each population.The final cluster assignment is shown by the colour underneath.(Bottom) A network of clusters indicating gene flow by weighted posterior admixture estimates for evaluated breeds.Direction of the gene flow is represented by arrows.