Submitted to: Journal of Animal Science
Publication Type: Abstract Only
Publication Acceptance Date: 5/1/2005
Publication Date: 9/1/2005
Citation: Blackburn, H.D., Welsh, C.S., Stewart, T. 2005. Within breed selection of boars for a gene bank. Journal of Animal Science. 83(Supl 1):138.
Interpretive Summary: A systematic approach for selecting animals for inclusion into a gene bank has received little attention. We report an approach that enables the selection of animals for such a purpose. This approach consists of utilizing pedigree information from breed associations. From the pedigree records genetic relationships can be calculated between all individuals. Then using the genetic relationships, clusters of related individuals can be made. Once the animals are clustered a strategy for picking animals within clusters was needed to select animals from a cluster for inclusion to the repository. Two selection approaches were compared to randomly selected boars. These were selecting 5 boars per cluster and weighting the number of boars selected per cluster based on the total number of boars in the cluster. Average genetic relationships of the boars selected by these three methods were compared and no difference was found. Therefore, either selection approach would be a satisfactory method to use as a within cluster sampling approach. In addition, the clustering approach helps guide the acquisition process because clusters that are not present in the repository are easily identified.
Technical Abstract: A primary component of genetic conservation of livestock species is the development of cryopreserved collections. Approaches to prioritizing breeds for conservation and entry into gene banks have been developed. But little attention has been given to determining how to select individuals within a breed. We addressed this issue by utilizing breed association pedigree records for Hampshire, Yorkshire and Duroc breeds. From these records genetic relationships were computed and utilized in a cluster analysis. The Ward Method of clustering, from SAS, was selected after testing several different approaches. For all breeds approximately 2,800 young boars were clustered, litter mates were excluded from the analysis. The pseudo t-test was used to determine where significant breaks in the clusters occurred. In addition, a pre-determined collection target of 100 boars per breed had been set. Significant divisions in the clusters occurred when approximately 20 clusters had been formed. Once the number of clusters were determined, average relationships within and between clusters were computed and inspected. Using the Yorkshire as an example, several approaches were evaluated for individual selection and compared to randomly selected boars (control). These approaches were: weighting the number selected by the number of animals in the cluster, and selecting 5 boars per cluster. Randomly selecting boars for each of these approaches the average relationships for the selected boars were: 0.08, 0.08, and 0.07 for control, weighted by cluster size and 5 boars/cluster, respectively. These results indicated that the two selection approaches would enable the selection of individual boars for a gene bank with an average relationship at or below the population average. In addition, the approach helps guide the acquisition process because clusters that are not present in the repository are easily identified.