Over the last decades, pedigree dog breeding has attracted public scrutiny over a number of welfare issues associated with conformation (e.g. Asher et al., 2009; Collins, L. M. et al., 2010; Collins et al., 2011), such as syringomyelia in Cavalier King Charles Spaniels and some other toy breeds (Rusbridge et al., 2000) or Brachycephalic Obstructive Airway Syndrome in French Bulldogs and other brachycephalic breeds (McGreevy, 2007), and the management of deleterious recessive conditions segregating in many breeds (Calboli et al., 2008; Mellersh, 2008; Summers et al., 2010; Leroy and Baumung, 2011). These issues have a complex background, with the practices which originally led to creation of breeds, i.e. inbreeding and breeding to a defined standard for physical and temperament characteristics, nowadays becoming a hindrance to the genetic health of the populations (McGreevy, 2007).
The problems in management of the gene pool in pedigree dog populations are easier to identify than to solve for two reasons: firstly, the population structure of most breeds is vastly different from typical pyramid structure observed in many livestock species, with many breeders contributing their animals to the breeding pool, and any individual breeder having at best only a weak control of the overall breeding population. Secondly, pedigree breeds are closed populations by definition, with many pedigree breeds having very small genetic bases. Web tools such as The Kennel Club’s Mate Select offer a means of sharing technical information on the genetic state of a breed and so can promote informed decision-making among all breeders within a breed towards beneficial outcomes. However, such outcomes rely on the tools being used and, to the authors’ knowledge, this study is the first to examine to what degree this is so. Mate Select is an important subject for such a study as it is one of few tools managed by a national kennel club, consolidating the complete records of pedigree dogs registered in the country, for decades. The tool offers information on 222 breeds with a wide range of census sizes, and collectively these breeds have over 230,000 registrations per year. The results show that the tool has a large community of users going beyond the UK, which has remained relatively constant in number since its launch, but usage can vary dramatically for breeds of similar census size, and is greater for those breeds with EBVs available.
The measures of usage and breed population size used in this study approximate the breeding population of a breed. Usage was assessed by the number of searches querying putative mating pairs as this represents the key decision point when the future gene pool is being shaped and the planning of matings is fundamental to the pedigree breeder. Not all of the bitches that were the subject of searches were bred from, i.e. not all became dams, however their presence in the search data represents one part of the breeder’s decision-making process. The measure of breed size was the numbers of bitches that produce registered offspring, i.e. are dams for the breed, as this most closely represents the size of breeding resources available for a breed. This choice avoids biases arising from variation in litter size, which would have influenced the total number of registrations for a breed, and complications associated with the mating ratios of dams to sires. With these definitions, the simple average across all breeds was 4.1 searches per dam over the period of the data and indicates significant use of Mate Select.
Ideally, for genetic management one might wish for more searches per bitch for the smaller breeds as choices are more critical when the genetic base is narrow, but this was not observed for the breeds with the smallest census sizes studied here. However, there are reasons why this expectation may be unreasonable. Firstly, when a breed has only a few dams there are also few dogs available, so the mate choices are limited by census size too. Secondly the bitches of those numerically small breeds may be owned by very few breeders and it is feasible that the information stored in Mate Select may also be maintained privately. Moreover, the UK ‘population’ may actually only be an outpost/subset of a larger (but still small) global breed population, which may have maintained pedigree databases independent of national registries. Beyond this group of the numerically smallest breeds, the expectation of more searches per dam for smaller breeds was evident, with a peak at 70 registered dams (see Figure 4b) followed by a decline as breed populations became larger. However, the variation around the trend line shown in Figure 4a shows there is considerable variation attributable to breed culture or structure among breeds of similar size. Among the extreme deviations, the reasons for the profoundly low search frequency for the Wire Fox Terrier remains a mystery, while the high search frequency for Norwegian Buhund has coincided with the release of a DNA test for cerebellar ataxia affecting this breed (Jenkins et al., 2020). While the last decades have seen a multitude of DNA tests developed for various breeds, no other coincidences of extreme search frequencies with emerging DNA tests were observed.
Some of the structural and cultural causes of variation among breed in searches per dam, after accounting for size, may prove to be predictable as demonstrated by the results for the EBV-track breeds in this study. These breeds have participated in BVA/Kennel Club recording schemes long enough to accumulate sufficient data for EBVs for hip and/or elbow dysplasia to be produced by the end of the study period. Participation in such schemes is itself a step towards collectively addressing a breed’s health problems, and some have consequently shown demonstrable improvement in the hip scores (James et al., 2020), and researching the phenotypic scores would encourage usage of Mate Select. EBVs first became available for some UK breeds in March 2014, midway through the study period, and the number of breeds and the scope of the EBVs for particular breeds has expanded at different times since then. When EBVs are produced for a breed, the estimates of genetic merit are more accurate and encompass the whole of the breed, not just those that have been phenotyped. Whilst the hypothesis that publication of the EBVs may be responsible for the increased searching on Mate Select for these breeds is plausible, the analysis in Supplementary Information 2 provides no support for this and shows that differences in usage existed prior to publication. The study has highlighted a number of factors that may be considered for increasing usage as tools are developed, beyond the importance of the providing good genetic content, such as EBVs, as a means of maintaining and promoting usage. Firstly, the rapid growth in demand for accessing the tools on mobile devices was very clear, and the trends shown in Figure 2 would predict that the majority of users would now be using Mate Select from mobile devices. This requires (i) the underlying software to be flexible enough to cope with the demand for providing information on different platforms, and (ii) as importantly, how the information is presented needs to account for this variation.Secondly, the tools can have global usage, although for Mate Select usage from outside the UK was relatively small (<5%) and primarily from within Europe and English-speaking countries. This too raises questions on presentation, not only in relation to language and terminology, but also on content. For example, EBVs currently provided for hip and elbow dysplasia within Mate Select are based on the scoring system developed by British Veterinary Association, UK. Thus, the EBVs are not directly comparable to EBVs developed in other countries and based on other scoring systems (e.g. Orthopedic Foundation for Animals in breeds registered with American KC, or Federation Cynologique Internationale in most other countries) (Flückiger, M., 2007). While combining the scores from different systems to calculate EBVs may not be feasible at this time, as it would require merging of worldwide pedigree databases, a useful first step would be to generate and present conversion formulae, which would be more meaningful to international users. Thirdly, usage has a clear cyclicity with periods of comparatively heavy use within the year. Awareness of this should be accounted for in the planning of down-times (e.g. for maintenance) and the launch of improvements to avoid discouraging usage in both the short- and long-term.