Statistical analysis for SNP

Accuracy of prediction for different numbers of SNPs

SNP chips are a relatively new technology that gives a snapshot of an individual’s genetics across the whole genome. AgResearch’s statisticians are working alongside scientists from the Animal Improvement section to harness this technology for the New Zealand sheep industry.

Until recently, animal selection using DNA had been restricted to a few well-characterised variants. The SNP chip technology provides an opportunity to select on many genes across the whole genome. This allows the genetic merit of an animal to be estimated without the need for information on their traits and pedigree.

AgResearch’s statisticians have helped develop the required prediction equations for use in the New Zealand sheep industry. This involved developing and applying quality control procedures, choosing a suitable prediction method, protocols for choosing animals to both ‘train’ the predictors and to validate them, ensuring the best genetic information is used while keeping training and validation information distinct, methods to accommodate population structure (e.g. breed mixes) and methods to estimate the accuracy of the predictor (how well the genetic merit is predicted).

The results of this work are commercially available through the Ovita consortium (owned jointly by Beef + Lamb New Zealand and AgResearch) and their commercialisation partner, Pfizer Animal Genetics.