Differential breeding practices drive subpopulation differentiation and contribute to the loss of genetic diversity within dog breed lineages

Background Discrete breed ideals are not restricted to delimiting dog breeds from another, but also are key drivers of subpopulation differentiation. As genetic differentiation due to population fragmentation results in increased rates of inbreeding and loss of genetic diversity, detecting and alleviating the reasons of population fragmentation can provide effective tools for the maintenance of healthy dog breeds. Results Using a genome wide SNP array, we detected genetic differentiation to subpopulations in six breeds, Belgian Shepherd, English Greyhound, Finnish Lapphund, Italian Greyhound, Labrador Retriever and Shetland Sheepdog, either due to geographical isolation or as a result of differential breeding strategies. The subpopulation differentiation was strongest in show dog lineages. Conclusions Besides geographical differentiation caused by founder effect and lack of gene flow, selection on champion looks or restricted pedigrees is a strong driver of population fragmentation. Artificial barriers for gene flow between the different subpopulations should be recognized and abolished for the maintenance of genetic diversity within a breed.


Background
Domestic dogs (Canis familiaris L.) are the oldest domesticated animals, comprising of more than 400 highly diverse contemporary breeds (1). The domestication process itself has caused genetic bottlenecks shaping dog development, the first dated to have occurred about 15,000 years ago (1,2). While early domestication was characterized by adaptation to the mutualistic relationship with humans, different dietary conditions and geographical differentiation, the recent bottlenecks are much more dramatic and caused by goaldirected breeding, especially during the last two centuries. In fact, the idea of discrete, uniform and standardized dog breeds originates from the Victorian era and current 3 breeding practices largely still reflect this thinking (3). Creation of breeds was driven by the advent of dog shows, where breed standards are contested and the best dogs in each class rewarded. Competition brings prestige and some income from sales and stud fees for the owners, but more importantly forces the specification of ideal breed conformation.
Establishment of a breed, by definition, causes effective cessation of gene flow to the population from other dog populations. Furthermore, extreme breed standardization as the main driver of breeding practices causes strong reduction of the effective population size and high levels of inbreeding within the breed, resulting in loss of genetic diversity and accumulation of deleterious alleles in many of the contemporary breeds (1,(4)(5)(6)(7). While extreme directional breeding is often resulting in exaggerated morphological or functional characteristics (8), also less specialized breeds -or so-called primitive breeds -are impacted by extreme breeding practices. We have recently shown that the excessive use of champion males has driven the loss of heterozygosity in the Finnish Spitz, in contrast to Nordic Spitz (9), which is a younger breed with more loosely defined standards. This is particularly interesting as the two breeds originate from the same feral founder population (10), making it possible to compare the outcomes of the different breeding practices in a similar genetic background.
While the impact of extreme breed standardization on dog population genetics is generally acknowledged, there can be differing breeding practices within breeds too. In gun dogs, selection on fur types and local breed sub types has often resulted in splitting of the breeds (11), but the extent of subpopulation differentiation within a breed due to diverse breeding practices is less known. In many of the contemporary breeds, there are existing divisions depending on the breeding goals as well as geography. If these divisions result in subpopulation differentiation, they could exacerbate the loss of genetic diversity within the breed. In the presented study, we sought to see whether the differential breeding 4 practices, such as partition to sport or show dogs, or geographical division has caused genetic differentiation into recognizable subpopulations in six popular dog breeds, Belgian Shepherd, English Greyhound, Italian Greyhound, Finnish Lapphund, Labrador Retriever and the Shetland Sheepdog.
Using a genome-wide survey of 1319 single nucleotide polymorphisms (SNPs), we show here that all of the studied breeds show subpopulation differentiation, either by their geographical origin, selection for performance or morphological traits. We conclude that the subpopulation differentiation should be taken into account in the breeding programs of the studied breeds to conserve their genetic diversity.

Results
We analyzed the genetic differentiation in Belgian Shepherd, English Greyhound, Italian Greyhound, Finnish Lapphund, Labrador Retriever and Shetland Sheepdog using multidimensional scaling (MDS) (12) and STRUCTURE (13) analysis of the genotypes of the 1319 SNPs. The observed population structure was matched with the ad hoc data on geographical location, use (sport vs. show dog) and morphological traits of the individual dogs (see methods). The example breeds were chosen as they represent either sport dogs used for competition (Greyhounds), are known to have subpopulations used for service or as companion dogs (Labrador Retriever), have differentially bred sub types (Belgian Shepherd), are locally (Finnish Lapphund) or globally (Shetland Sheepdog) popular companion dogs.
In contrast to our previous study on Nordic hunting dogs (10), all of the breeds studied here showed at least some level of subpopulation differentiation. Italian Greyhound and the Shetland Sheepdog show clear geographical subdivision with the dogs from different continents clustering together ( Figure 1). It is noteworthy that although STRUCTURE analysis supported ancestral population differentiation among the European dogs in both breeds, as indicated by the different colours ( Figure 1B and D), there seems to be relatively frequent admixture reflecting the exchange and import of breeding dogs between the European countries. The fact that the US populations of both breeds are strikingly different from their European counterparts is likely mainly due to a founder effect, resulting in higher F ST and lower heterozygosity (Hz) values (Table 1) Table   1).
The Belgian Shepherd Dog has four different varieties: shorthaired Malinois, wirehaired Laekenois, and longhaired Tervueren and Groenendael. Tervueren and Groenendael differ by their coat colour, the former being sable and the latter black. In the US, the name "Belgian Sheepdog" is reserved for and used when referring to Groenendaels. The US originating "Belgian Sheepdogs" were omitted from the analysis as, although clustering with the European Groenendael, they also showed population differentiation caused by founder effect. As expected, the Malinois population consisting mostly of working dogs, forms its own distinct cluster ( Figure 3A). Rather unexpectedly, this cluster also contains the sampled Laekenois; a finding also supported by the CLUSTER analysis ( Figure 3C).
Despite their coat type difference, both breed forms are used as working dogs and crossbreeding of the two types is allowed, at least in Finland, where the Laekenois specimens originated from. The longhaired Groenendael and Tervueren were more heterogenic, with some Tervuerens being indistinguishable from Groenendaels.
The last breed in our analysis was the Lapphund, a fairly young breed, originating from the Sámi reindeer herding dogs together with the Lapponian herder (10). In contrast to the other primitive spitz-type Nordic breeds, including its sister breed the Lapponian herder The speed of the differentiation is remarkable, as the association following the restricted breeding practises was founded in 1981.

Discussion 7
In general, the efficacy of population genetic studies is dependent on the sample size as well as on the number and type of the chosen markers. Although dwarfed by the existing 170k arrays for dogs (6,14,15), the 1319 diagnostic neutral SNPs used in the study have previously been successfully used in detecting dog population structure as well as inferring their genetic relationships (9,10). In more detail, as the distance between the markers is known, the 1319 SNPs were successfully used for an r 2 based estimation of the historical effective population sizes of the Finnish and Nordic Spitz (9). In general, analysis of SNPs across all chromosomes can outperform any traditionally used STR panels in population genetic analyses (16).
Our analysis was able to demonstrate subpopulation differentiation in all of the studied six breeds and confirmed that both geographical isolation as well as differential breeding strategies can have similar outcomes and that the differentiation can be fairly rapid, as seen in the Finnish Lapphund. If the reproductive isolation was to be complemented with strong directional selection, we would expect the differentiation to be even faster and more extreme. It is noteworthy that the differentiated subpopulations having high F ST values (Table 1), formed tight clusters on the MDS plots and were more uniform in their STRUCTURE charts than the less differentiated examples (Figures 1-3).
The geographic differentiation between the European and the US populations of the Italian Greyhounds and the Shetland Sheepdogs was expected. Both breeds originate from Europe and the US populations are based on few founders (17). The founder effect combined with the limited gene flow between the continents is expected to be efficient drivers of subpopulation differentiation. What is more astounding is that similar or even more extreme differentiation rates, as evident in the F ST values (Table 1) Population fragmentation is an issue for the conservation of the genetic diversity in any population. The measure of population differentiation F ST , by definition, is in effect a measure of inbreeding in the subpopulation relative to the total population (18,19).
Inbreeding itself reduces the heterozygosity in a population and in fact inbreeding can be expressed also as a function of loss of Hz over generations (see (9) for the different metrics and discussion in dogs). This is also evident in our study, where the subpopulations with the highest F ST had also the lowest Hz (Table 1). If the obstacle for the gene flow is maintained, Hz will continue to erode at higher rates than in the less differentiated populations.
The loss of genetic diversity can pose a threat to breed health (17,(20)(21)(22)(23)(24). Although this loss cannot be completely avoided due to the closed population nature of all dog breeds, some breeding practices, such as extreme selection for "best in breed" winners of competitions and trials or lineages with culturally valued pedigrees have the potential to accelerate the process. Knowledge on the genetic differentiation of populations and gene flow between them not only offers insights into breed history and current breeding practices, but also highlights how subpopulations might most beneficially be used for the maintenance or restoration of genetic diversity.

Conclusions
Our results show that geographical isolation and founder effect are the main drivers of subpopulation differentiation among Italian Greyhounds and Shetland Sheepdogs, whereas selection for show and sports or working lineages explained the genetic structure among English Greyhounds and Labrador Retrievers. The genotyping analysis could also detect the predicted breed type structure in Belgian Shepherds and revealed unexpected, relatively recent split in the Finnish Lapphund population due to breeder preferences. Our study exemplifies the use of genetic data in detecting population structure among dogs to identify and measure the degree of subpopulation differentiation. The findings could be then applied in breeding programs to facilitate gene flow between isolated subpopulations, as well as to raise awareness of practices that are potentially harmful for the maintenance of genetic diversity within a breed.

Breed selection and DNA sampling. The Belgian Shepherd, English Greyhound, Italian
Greyhound, Labrador Retriever, Finnish Lapphund and Shetland Sheepdog breeds were selected for the study as they were known or suspected to have either distinct breed varieties (Belgian Shepherd) or differential breeding practices depending on the intended use of the dogs ( Table 2). Breeds without expected subpopulations were not included in this study, as we have recently analysed several Nordic hunting dog breeds using the same methods as a proof-of-concept, and shown that these breeds exhibited no subpopulation division despite evidence of ancestral admixture (10). Genetic analyses were carried out on DNA extracted from owner-collected, non-invasive buccal swab samples. All dog owners provided consent for the use of their dog's DNA sample for research purposes.
Genotyping. The genotypes were obtained using a custom-designed Illumina Infinium (Illumina, Inc., San Diego, CA, USA.) genotyping microarray for 1319 neutral SNPs, evenly distributed across the 39 canine chromosome pairs (25) with a median intermarker distance of 1,585 kilobases. The array was previously shown to reliably differentiate between breeds, detect population structure and trace their ancestry (10). The spacing of the markers allows also the estimation of the effective population sizes based on the decay of linkage disequilibrium (9). Genotypes for all samples are provided as Additional