Retrospective analysis for genetic improvement of hip joints of cohort labrador retrievers in the United States: 1970-2007

Abstract

Background: Canine Hip Dysplasia (CHD) is a common inherited disease that affects dog wellbeing and causes a heavy financial and emotional burden to dog owners and breeders due to secondary hip osteoarthritis. The Orthopedic Foundation for Animals (OFA) initiated a program in the 1960's to radiograph hip and elbow joints and release the OFA scores to the public for breeding dogs against CHD. Over last four decades, more than one million radiographic scores have been released.

Methodology/principal findings: The pedigrees in the OFA database consisted of 258,851 Labrador retrievers, the major breed scored by the OFA (25% of total records). Of these, 154,352 dogs had an OFA hip score reported between 1970 and 2007. The rest of the dogs (104,499) were the ancestors of the 154,352 dogs to link the pedigree relationships. The OFA hip score is based on a 7-point scale with the best ranked as 1 (excellent) and the worst hip dysplasia as 7. A mixed linear model was used to estimate the effects of age, sex, and test year period and to predict the breeding value for each dog. Additive genetic and residual variances were estimated using the average information restricted maximum likelihood procedure. The analysis also provided an inbreeding coefficient for each dog. The hip scores averaged 1.93 (+/-SD = 0.59) and the heritability was 0.21. A steady genetic improvement has accrued over the four decades. The breeding values decreased (improved) linearly. By the end of 2005, the total genetic improvement was 0.1 units, which is equivalent to 17% of the total phenotypic standard deviation.

Conclusion/significance: A steady genetic improvement has been achieved through the selection based on the raw phenotype released by the OFA. As the heritability of the hip score was on the low end (0.21) of reported ranges, we propose that selection based on breeding values will result in more rapid genetic improvement than breeding based on phenotypic selection alone.

Figure 1. Distribution of dogs born and scored between 1970 and 2007.

There were 154,352 Labrador retrievers scored by the Orthopedic Foundation for Animals (OFA) during this period. The number of dogs are indicated by the blue bars for the year that the dogs were born and by the red bars for the year that the dogs were scored. Both the minimum and the median age of scoring were two years old. The lag between the birth year and scoring year was over two years.

Figure 2. The phenotypic and genetic trends in last four decades.

The trends are presented by the mean (± standard error) within each year for Orthopedic Foundation for Animals (OFA) score (the vertical axis on the left) and breeding value (the vertical axis on the right) for the 258,851 Labrador retrievers born between 1970 and 2005.

Figure 3. The joint and marginal distribution of breeding value and accuracy.

Breeding value (BV) is displayed on the vertical axis and its accuracy is plotted on the horizontal axis for the 258,851 Labrador retrievers born between 1970 and 2005. The marginal distributions of BV and accuracy are indicated by both the total number of dogs and percentile at each category. The dogs on the upper right with lower BV (better hip) and higher accuracy are the most ideal for breeding against hip dysplasia.

Figure 4. Trend of inbreeding coefficients.

The trend is presented by the mean (± standard error) within each year for the 258,851 Labrador retrievers born between 1970 and 2005.