Genetic evaluation of guide dogs in the UK

Guide Dogs is the largest assistance dogs charity providing mobility support for people who are blind or partially sighted. Approximately 70% of dogs bred become working guide dogs, and this value is high compared with other guide dog schools. However, it is still desirable to increase further the proportion of dogs that are bred which successfully become guide dogs. This study aimed to complete a thorough genetic evaluation of Guide Dogs’ programme and colony to investigate the potential for using quantitative genetic tools in service of this aim.
\n
\nFirstly, reasons for withdrawal of dogs from Guide Dogs’ programme between 1995 and 2012 were analysed, and a survey of selection aims was undertaken among the seven individuals involved with selecting breeding stock, to ascertain which health and behavioural traits were of most importance. Health and behavioural traits were approximately equally weighted in the survey, but behavioural reasons accounted for 68% of withdrawals of dogs from Guide Dogs’ programme. A key finding of the survey was that selection aims were breed-specific.
\n
\nGenetic evaluation of health and behavioural traits recorded by Guide Dogs was then undertaken. Historical health records were interrogated and cases of disease conditions were collated. Heritability analyses were conducted and genetic correlations between disease conditions were investigated. Atopic dermatitis, cranial cruciate ligament (CCL) disease, diabetes mellitus, distichiasis, elbow dysplasia, entropion, hip dysplasia, laryngeal paralysis, multifocal retinal dysplasia, panosteitis, patellar luxation and seizures were all found to be heritable in Labrador Retrievers. High, positive genetic correlations were found in this breed between hip and elbow dysplasia, hip and elbow dysplasia and panosteitis, and elbow dysplasia and seizures. Heritability estimates were reported for atopic dermatitis, congenital ichthyosis, entropion, Horner’s syndrome and panosteitis in Golden Retrievers. In German Shepherd Dogs atopic dermatitis, hip dysplasia, panosteitis and sebaceous cysts were found to be heritable. Most heritability estimates were small or moderate in magnitude. Selective breeding strategies that identify those animals with low genetic risk, such as the use of estimated breeding values (EBVs), could be used to reduce the incidence of these conditions.
\n
\nCrossbreeding parameters were investigated in Labrador Retrievers, Golden Retrievers and crosses between the two breeds for atopic dermatitis, CCL disease, elbow dysplasia, entropion, hip dysplasia, Horner’s syndrome, panosteitis and seizures. Heterosis appeared to slightly reduce the likelihood of developing elbow dysplasia, Horner’s syndrome and seizures. Recombination loss appeared to slightly increase the likelihood of developing either hip or elbow dysplasia. Increasing Labrador fraction was associated with a greater probability of developing elbow dysplasia but a lower probability of developing Horner’s syndrome. These results suggest that there are small benefits of heterosis for the first generation (F1) cross but that these may be lost when the F1 is crossed back to a purebred dog.
\n
\nCurrent Guide Dogs practice was to “measure” behavioural traits using two different behavioural assessments. Results from these assessments were also subjected to genetic evaluation. Firstly genetic and environmental parameters relating to scores in the Canine Assessment Summary (CAS) were estimated in purebred Labrador Retrievers and Golden Retrievers and then in crossbreed models between the two breeds and their crosses. Although many of the models measured heritability estimates which were not detectably larger than zero, and the assessor effect was considerably larger than the heritability estimate in most cases, the heritability estimates suggest that there is substantial genetic variation in many of the traits being measured by CAS. Therefore EBVs for these traits could be used, with scores at the first CAS assessment in advanced training looking particularly suitable. Bivariate models in Labrador Retrievers identified three negative genetic correlations between behavioural traits: between confidence and distraction, eagerness and interaction with people and calmness and eagerness. These could be problematic as selection for one of these traits could lead to a worsening in the other.
\n
\nMost of the crossbreeding parameter estimates were not detectably larger than zero. However, seven CAS elements had small to moderate heterosis estimates, six of which were negative i.e. beneficial. Four CAS elements had small to moderate estimates of recombination loss, three of which were positive i.e. detrimental. This suggests that, as with the disease conditions, there may be benefits in behavioural terms in the F1 but these may be at least partially lost in the backcrosses. Repeatability models were also undertaken for two CAS elements, calmness and eagerness. For both traits estimates of permanent environmental effects were larger than heritability estimates suggesting that they are more important than genetic influences on these traits.
\n
\nFinally heritability and crossbreeding parameters were estimated for scores in the Puppy Profiling Assessment (PPA). Nine of the 11 PPA components had heritability estimates which were detectably larger than zero, five of which were moderate in magnitude, indicating that performance in these tests had an inherited element. Crossbreeding parameter estimates for PPA components were mostly not detectably larger than zero however; the PPA dataset may have lacked the power to detect crossbreeding effects due to its relatively small size.
\n
\nThis study provides a platform for the implementation of quantitative genetic techniques to improve the accuracy of Guide Dogs’ selection decisions. Many of the findings of the study will also be of interest to the wider dog breeding community.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.