#40 · Genomics

GxE-aware genomic prediction improves Miscanthus biomass yield forecasts

The three preprints converge on genomic prediction for Miscanthus breeding populations evaluated across multiple environments. Their common thread is explicit modeling of genotype-by-environment structure rather than reliance on main-effect models. One study reports that GxE-aware models raise predictive ability over main-effect models across cross-validation schemes for biomass yield. A second evaluates sparse testing designs for multi-environment trials, finding that the GxE model performs best and that phenotyping costs could be cut about fivefold. A third benchmarks five genomic prediction models, including multi-trait multi-environment approaches, on 336 M. sacchariflorus genotypes under three cross-validation schemes. What is new is the joint focus on testing-design economy and model structure in this crop.