Evaluation of half-sib families to address yield decline in kernza intermediate wheatgrass by assessing genomic prediction and genetic gain

Authors: Bajgain, Prabin; JUNGERS, JACOB - University Of Minnesota; ANDERSON, JAMES - University Of Minnesota

Submitted to: Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 9/23/2025
Publication Date: 11/10/2025
Citation: Bajgain, P., Jungers, J.M., Anderson, J.A. 2025. Evaluation of half-sib families to address yield decline in kernza intermediate wheatgrass by assessing genomic prediction and genetic gain. Proceedings. https://scisoc.confex.com/scisoc/2025am/meetingapp.cgi/Paper/168833.

Interpretive Summary: Intermediate wheatgrass breeding programs mostly assess traits on hundreds of individual plants. Farmers, on the other hand, establish plants in closely arranged rows for large-scale grain production. A breeder's plots are thus not always representative of large commercial fields. In this experiment, we established plots in rows, as a farmer would, and measured plant performance over multiple locations and years to identify plants with stable yield as they aged. We identified genetic regions controlling important traits such as yield, seed size, and height. With the collected data, we also trained genomic selection models to predict trait performance, and assessed if trait prediction can reliably be done in populations established in row-plots. Based on our findings, we recommend that row plots be evaluated regularly by breeding programs for selection and genetic improvement. Doing so will benefit growers by making the data more relatable to their farming operations.

Technical Abstract: The perennial forage grass Thinopyrum intermedium (intermediate wheatgrass, IWG, Kernza®) has been domesticated by the University of Minnesota since 2011 to exploit its extensive roots, high biomass, and potential to produce grain. Our breeding program has relied on assessment of individual spaced-plant performance to train genomic prediction models and select parents for germplasm improvement and construction of synthetic cultivar candidates. This common methodology does not reflect the densely planted rows farmers use in large-scale grain production. Recognizing that spaced-plant nurseries differ from commercial row-planted stands, we assessed 208 half-sib families (HSFs) in sward plots across two Minnesota sites over three years, measuring plant height, grain yield, and seed size. Substantial trait variation was observed across environments and 32 HSFs maintained stable yield across sites and years, countering yield decline observations that are typical beginning in the 3rd growing season. Broad-sense heritabilities ranged from 0.36 (yield) to 0.51 (height), and genome scans uncovered 42 marker–trait associations. Simultaneously, comparisons of univariate and multivariate genomic prediction models showed that multivariate approaches outperformed univariate ones by up to 49 percentage points, although no single model excelled for all traits, and cross-environment accuracy peaked in St. Paul. Genetic gain estimates suggested theoretical improvements of 11 g grain yield and a 3 cm reduction in height per cycle, underscoring that integrating row-plot evaluation of HSFs with genomic predictions can accelerate IWG breeding. Based on our findings, we recommend that row plots be evaluated regularly by IWG breeding programs for genotype selection and genetic improvement of breeding populations. Doing so will benefit Kernza growers by making the data collected in breeding trials more relatable to their farming operations.