Easily generate insights from skin microbiome samples with MORPHEUS
The cultivation of malting barley is strongly challenged by changing climate conditions, and a need has emerged for cultivars that deliver stable yield in multiple environments while maintaining excellent quality for malting. The goal is to breed barley varieties by applying advanced breeding predictions that leverage all available data on genotypes, phenotypes (yield + malting quality), environments and climates.
This helps to secure AB InBev’s position as a leader in the beverage market with access to the highest quality malting barley varieties.
Computomics used xSeedScore to develop an advanced model for genomic predictions of barley crosses that included not only genotyping and phenotyping data, but also information on environments from climate and drone data.
Due to Computomics’ advanced machine learning technology, the model is able to learn and predict non-linear effects on multiple traits. Our computational power allows the simulation of millions of potential crosses up to the F4 to F6 generation with their complex multi-trait phenotypes. From these, AB InBev’s breeders choose the most promising crosses based on yield and multiple quality traits to advance into the next stage. Each breeding cycle predictions are validated against field results and the model is optimized to further increase its prediction accuracy.
Compared to BLUP models, this method takes complementing markers into account so that it is possible to predict if offspring phenotypes exceed those of their parents.
All results are visualized for easier and faster interpretation.
The use of xSeedScore allowed AB InBev to choose the best parents to advance their barley breeding programs by selecting from the broadest range of potential crosses. This approach reduced the risk of missing good combinations and speeds up the development of commercially successful lines. Furthermore, our method reduced the number of tests for malting by integrating early stage predictions of malting traits.
The integration of environmental data allows us to develop future-proof varieties for specific locations and a changing climate.
Predict virtual hybrids from a male and female double-haploid population and predict hybrid phenotypes that exceed their parents and testers