Host: Professor Brande Wulff

Abstract:
The rapid expansion of genome-wide data is driving biological discovery and enabling applications such as selective breeding in agriculture. However, as data volume increases, storage becomes a challenge, and traditional genome-wide statistical models struggle to fully harness this wealth of information for deeper insights and improved applications. In this talk, I will present our work on leveraging ancestral recombination graphs (ARGs) encoded in the tree sequence data format. We focus on ARGs for three key reasons: (1) efficient storage of large-scale whole-genome sequence data, (2) improved understanding of genomic diversity across populations, and (3) the development of a richer statistical model for quantitative genomics that fully capture branching/coalescence, mutation, and recombination events. I will illustrate our results from analyzing the 1000 Bull Genomes Project, data from Japonica and Indica rice varieties in a breeding program in Uruguay, and simulations. The later show that we can scale the implicit ARG covariance matrix times a vector product to millions of genomes.

Bio:
Gregor Gorjanc is a professor of Selective Breeding at The Roslin Institute, University of Edinburgh, where he also holds affiliations with the Global Academy for Agriculture and Food Systems, and the Centre for Statistics. He earned his engineering degree in Animal Science and a Ph.D. in Genetics from the Biotechnical Faculty at the University of Ljubljana in Slovenia. 

Gorjanc leads the HighlanderLab, focusing on managing and improving populations using data science, genetics, and breeding. The lab focuses on populations used for the production of food, feed, and fibre. Their research interests include (i) developing methods for genetics and breeding, (ii) designing and optimizing breeding programs, and (iii) analyzing data to uncover biological insights and innovative approaches to population improvement.