Genetic Diversity of Great Dane Breed Using Ten Microsatellites : Impact of Breeding Control over the Breeding Line

The American Kennel Club and the Fédération Cynologique Internationale recognize two phenotypic variants of Great Dane breed denominated American and European varieties. Historically, these varieties have been segregated according to morphological characteristics. In an attempt to obtain a better characterization, breeders have been interested in the genetic parameters that could evaluate the within-breed diversity. In this document, we studied the genetic structure of Great Dane breed with 10 STR markers in 88 dogs using capillary electrophoresis. Cluster analysis, population differentiation and phylogenetic analyses revealed that American and European varieties are genetically independent. Nevertheless, within the American variety a genetic stratification was found. Additionally, a high misclassification (28%) was detected, which could be due to wrong registration or false paternity. Our results support the importance to deem genetic markers as useful tools in breeding control. Similarly, these studies serve as an accurate reference to establish standards by dog breeding associations and for choosing among dogs for inter-breeding. Nevertheless, genetic tools are only a complement of morphological methods, since both are reshuffling These authors equally contributed to this work. Corresponding author.


Introduction
In agreement with the American Kennel Club (AKC) and the Fédération Cynologique Internationale (FCI) there are two recognized varieties of the Great Dane breed: American and European [1].The differences between these varieties are exclusively associated with morphologic characteristics.However, with the advent of new bioinformatics tools, Kennel clubs have become more interested in genetic studies to evaluate the within-breed diversity.These studies are a powerful tool to obtain a better characterization by combining both genetic and phenotypic characteristics [2].Short tandem repeats (STRs) are the preferred markers to study genetic variability in a population as well as parentage verification, suggesting that different breeds display distinct allele frequencies.Hence, STRs provide information about the genetic composition of purebred canine populations [3]- [7].In this light, recent studies have demonstrated considerable genetic differentiation even within recognized breeds [8].These differences could be related to artificial selection, demographic events (bottleneck and effective population size), as well as population stratification.In an attempt to gain more insight into these differences, an increasing number of studies have been focused on characterizing genetic structure among and within canine populations.Nevertheless, little information is available and analysis of genetic structure in local populations is still limited.
In order to study genetic variation in Great Danes, we characterized within-breed genetic differences.We analyzed 10 STR markers in 176 chromosomes using capillary electrophoresis.The population genetic structure, population differentiation, and phylogenetic analyses showed that American and European varieties are genetically independent.In addition, cluster analyses using Bayesian methods suggested that the American variety is stratified in two subpopulations, one of them showing an excess of heterozygosity (F = −0.0600).This data indicate that crossbreeding between different varieties of Great Dane dogs, to correct for genetic diseases, has originated genetic stratification within the American variety of Great Danes.Our results also show that this approach is useful for addressing dog breed genetic differences, even among varieties within the same breed, and corroborates the importance of determining the within-breed genetic structure in order to maintain breeding lines.

Population Studied
In cooperation with the Herrán-Aguirre kennel and the Mexican Great Dane Club, 88 dog records were compiled from animals with breeding license.From each individual we obtained information regarding origin, birthplace, age and sex by reviewing individual registration (pedigree) data provided by the owners.Great Dane dogs were phenotypically classified as American (65 dogs) or European (23 dogs), following the American Kennel Club (AKC) and the Fédération Cynologique Internationale (FCI) standards (Table 1), and based on the reiterated information on the dogs' pedigree.

Sampling
Blood samples (4 mL) were collected using the vacutainer system, with EDTA as anticoagulant.DNA was isolated from these samples and purified using the Qiamp DNA Mini Kit (Qiagen, Düsseldorf, Germany), following the manufacturer's instructions.
To infer population structure, we used principal component analysis with the GenAlEx v. 6.41 [12] software followed by a model based on Bayesian statistics using the Structure v. 2.3X software [13].Briefly, the genetic background of each individual is represented by a vector, 1 , , k q q q =  , where q k is the proportion of the indi- vidual genome originated in the population k.The software uses the Markov Chain Monte Carlo (MCMC) method to estimate the number of populations, allele frequencies in each population, and the q value for each sampled individual.All runs were performed with 100,000 length of burn-in period and 700,000 number of MCMC repeats after burn-in.This procedure calculates the probability that the observed set of genotypes could occur across a designated range of possible k's.
Allele frequencies of each population were used to construct a dendrogram.The analysis was done using the Phylip package v. 3.69 [14] and Treeview v. 1.6.6 software [15].The Neighbor-Joining method (NJ) was used to display the matrix of pairwise distance to compute Reynold's standard genetic distances among the different populations.Also, a bootstrap method (100,000 replications) was used to test the significance of the population cluster in the dendrogram [16]- [18].

Population Genetic Structure
Two varieties of Great Dane were defined by phenotypic characteristics (Table 1).The 88 dogs were originally classified by phenotypic parameters as 65 Americans and 23 Europeans.
To establish whether these varieties showed genetic differences, we determined the genetic structure of the Great Dane breed using principal component analysis (PCA) as a preliminary method.As a result, we obtained three different Great Dane clusters, which can be considered as three independent genetic populations (data not shown).In order to confirm these findings, we used a Bayesian clustering method to estimate the number of distinct populations and calculated the within-breed genetic diversity.We analyzed 88 individuals and found that they were clustered into three different genetic populations (k = 3, P = 1).The individuals were sorted as follows: by dendrogram construction (Figure 2).
As a result of the genetic analysis, we found that of the 65 dogs that were originally classified as Americans based on phenotype, only 47 (72%) were genetically Americans, and 18 (28%) were genetically reclassified as Europeans.Alternatively, of the 23 phenotypically European dogs, 18 (78%) remained as genetically Europeans, but 5 (22%) were genetically reclassified as Americans.
In addition, a population differentiation analysis was performed to detect genetic differences among the defined groups (Table 2).In this regard, important differences were found between subpopulations 2 and 3 (P <   0.0001 ± 0.0000), as well as between subpopulations 1 and 3 (P = 0.00450 ± 0.0019).On the other hand, when subpopulations 1 and subpopulation 2 were compared, only 5 out of 10 loci showed significant differences (P < 0.05), including the more informative loci.
Ultimately, the inbreeding coefficients (F) for each variety using the 10 loci were calculated.The lowest F value was observed in subpopulation 2 (F = −0.0600),which indicates an excess of heterozygosity, whereas subpopulation 1 and subpopulation 3 showed values of F = 0.0450 and F = 0.0150, respectively, with the consequent homozygote excess.

Discussion
This study comprises genetic information concerning the recognized Great Dane varieties, in which we showed the high efficiency of genetic markers in the control over the breeding line.
Our results reveal that the two recognized varieties of the Great Dane breed are genetically different, as has been assumed by its phenotypic characteristics.Nevertheless, cluster analysis confirms three different genetic subpopulations in the Great Dane population studied.Both subpopulation 1 and subpopulation 3exhibited an excess of homozigosity, which is related to inbreeding as a consequence of pure lineages.Based on phenotypic parameters, these populations correspond to the American and European varieties, respectively.Conversely, the third subpopulation showed an increase in heterozigosity levels (F = −0.0600),which could be associated with recent mixtures between dogs that are not of the same variety.Consequently, subpopulation 3 may be the result of sub-structural differentiation also named "breed split" caused by artificial selection.This non-natural selection carried out by breeders segregates varieties not only phenotypically, but genetically [8].
As a consequence of the lack of control in the genetic diversity within-breed, we found an important amount of discrepancies in classification.In light of this evidence, 25% of the dogs were reclassified when using genetic bases.These discrepancies could be related to wrong registration as well as false paternity.In this regard, previous studies have reported that dog breeds showed around 10% of average pedigree error, whereas the frequency of false paternity was close to 5% [19]- [21].In addition, the high misclassification could be related to crossbreeding between pure lines and recent admixture dogs.Although this miscegenation maintains morphological characteristics, the genetic imprint is modified.This practice may increase the European genetic lineage, though preserving the American phenotype, given that both varieties share an European ancestor.As a result of these mating practices, we found that 28% of the dogs with American pedigree were genetically classified as Europeans.Our findings are in agreement with recent studies, which support up to 30% of genetic variation in pure breeds [22].This genetic variation can cause important genetic differences even among different countries as Quingnon and colleagues point out [23].
Our results provide support to deem genetic markers as useful tools in breeding control to increase the quality in the breeding line [2].These methods can also help in health control and even in preserving athletic attributes, which have been gaining steady ground in the Canidae family [3] [24]- [26].Both genetic and morphological characteristics are important features to determine dog breed or variety.Nevertheless, STR genotyping should not be considered unique in defining the variety of an individual because this would lead to inaccuracies [7] [27]- [29].Although our data evinces the complex genetic architecture in within-breed dogs, it is advisable to increase the population size to discard sampling artifacts.In addition, further research is needed in which the number of STRs is increased to identify additional genetic characteristics.

Conclusion
Our findings demonstrate the usefulness of microsatellites in breeding and selection practices.Concomitantly, it is important to evaluate the genetic characteristics within breeds to maintain the breeding lines.Besides, our results support the use of STRs as a powerful tool to assign each variety rightly.Finally, it is also noteworthy to mention the usefulness of this type of studies to reduce genetic diseases by the identification of risk factors.To our knowledge, this is the first report of genetic diversity in Great Dane breed.

Figure 1 .
Figure 1.Genetic structure of the Great Dane breed.(a) Triangle pigureEb population 3) (Dane breed lost showing the genetic structure.(b) The bar plot represents the amount of genome of each individual that originated from each of three subpopulations represented as red (subpopulation 1), green (subpopulation 2), and blue (subpopulation 3).

Table 1 .
Phenotypic differences between American and European varieties.

Table 2 .
Population differentiation among the defined groups.

Table 3 .
Allelic frequencies and descriptive statistics parameters regarding the 10 STR loci in the three cluster found in Great Dane.Inbreeding coefficient.Bold numbers mean private alleles.