Hybrid StatsPD@Waite meeting

Every month, the professional development meetings of statisticians and data scientists at Waite, known as StatsPD@Waite, bring together specialists in various aspects of data sciences in agriculture from Waite, Roseworthy and Adelaide.

Please join us for the next StatsPD@Waite seminar where Daniel Tolhurst from Edinburgh Innovations, The Roslin Institute, Royal (Dick) School of Veterinary Studies, University of Edinburgh, UK will present on “Leveraging genotype by environment interaction for simulation and selection in plant breeding”.

Also please note that the StatsPD@Waite meetings are recorded. If you have a question to the speaker but had rather not be recorded, please send your question via chat during the meeting and it will be asked on your behalf.

Please email Beata Sznajder with questions or for details of the Zoom meeting.

Title: Leveraging genotype by environment interaction for simulation and selection in plant breeding

Presenter:Daniel Tolhurst, Edinburgh Innovations | The Roslin Institute | Royal (Dick) School of Veterinary Studies, University of Edinburgh, UK

Plant breeding is complicated by the fact that individuals (genotypes) respond differently to different environments, a phenomenon known as genotype by environment interaction (GEI). Despite its importance, however, many plant breeding programmes still use inefficient methods to handle GEI. This talk will cover new methods that leverage GEI for efficient simulation and selection in plant breeding.

The methods are built on multiplicative models, which are effective at capturing GEI with regards to a small number of multiplicative terms. This feature makes multiplicative models well suited to simulating GEI, whereby realistic structure can be generated between environments and genotypes with little computational effort. A new framework is presented for simulating large-scale plant breeding data.

Multiplicative models are also well suited for selection purposes because they are effective at disentangling non-crossover and crossover GEI. The premise here is that measures of overall genotype performance should be based exclusively on non-crossover GEI, whereas measures of stability should be based exclusively on crossover GEI. A new approach is presented that guarantees these features.

The methods are demonstrated using simulated data as well as collaborating cotton breeding data from Bayer CropScience.