genomic prediction

We use mixed model equations to create genomic predictions for economically important traits in beef cattle. Two of the limitations of application of genomic prediction in beef cattle are the price of the assay and the accuracy of the prediction. Our research aims to overcome these limitations.

population structure of cattle breeds

We use various computational methods to investigate the genetic histories of world-wide cattle breeds. We are interested in building the family tree of cattle breeds, as well as understanding the domestication of cattle.

identifying loci responding to selection

In 2012 we published a method, called Birth Date Selection Mapping, to identify loci responding to current selection. In this analysis we fit birth date (as a surrogate to generation number) as the dependent variable in a mixed model equation. Variants that have changed in frequency rapidly due to selection are strongly predictive of birth date, thus the method identifies regions under selection. The mixed model equations correct for relatedness and population structure within the data.

We have previously used this method in Angus cattle using approximately 45,000 SNPs. We are currently using the method in 5 cattle breeds using approximately 400,000 SNPs.

We are also interested in applying this method in other species.

population genetics

We also have various other population genetic collaborations, such as in Brassica and mad toms.