Loading…
Wednesday September 25, 2024 3:00pm - 3:15pm HST
Multi-location genomic prediction models have been proposed as a strategy for aggregating disconnected phenotypic datasets to enable performance prediction of individuals into new environments. This approach also offers improved prediction accuracy and understanding of environmental stability. Multi-location GBLUP prediction models have been developed and used to study patterns in genotype-by-environment interaction for: flowering date in cherry across four locations in USA, France, and Italy; for sweetness in peach across four locations in USA; in strawberry across eight locations in USA, Australia, and Europe; and for fruit firmness in apple across five locations in USA and Europe. Here we present an online bioinformatics platform for the deployment of these models via the Breeding Information Management System (BIMS) in the Genome Database for Rosaceae (GDR). This platform enables performance of new material that has not been physically tested to be predicted for the traits across the environmental space included in the multi-environment GBLUP models. The predicted genetic value of each individual included in the GBLUP models for each trait in each environment was used with the GBLUP individual-by-marker allele frequency matrix to predict the genetic effect of each marker allele for each trait in each environment. A portal was then developed on BIMS to host this vector of marker effects and enable new users to upload tables of new individual-by-SNP marker data for the loci used in the original GBLUP model. Predictions for the new individuals are produced by multiplying the new sample-by-marker frequency matrix by the vector of predicted allelic effects. Predictions of the new material for each trait across environments are delivered alongside predicted performance of publicly available genetic material to enable comparison and evaluation of reliability of predictions. The use of this portal is demonstrated with genome-wide SNP genotypic data collected on new apple selections evaluated for an Australian genetic improvement program.
Speakers
CH

Craig Hardner

University Of Queensland
Assocaite Professor in quantitative genetic s and tree breeding of horticultural and forestry crops with particular foci on multi-variate mixed linear models, G-by-E, genomic prediction, multi-trait selection, modelling breeding strategies, conservation genetics, low-cost genotyping... Read More →
Co-authors
CP

Cameron Peace

Washington State University
DM

Dorrie Main

Washington State University
SG

Shashi Goonetilleke

University Of Queensland
SJ

Sook Jung

Washington State University
NA
TL

Taein Lee

Washington State University
NA
Wednesday September 25, 2024 3:00pm - 3:15pm HST
Nautilus 1

Attendees (2)


Log in to save this to your schedule, view media, leave feedback and see who's attending!

Share Modal

Share this link via

Or copy link