Predictor bias in genomic and phenomic selection

• DOI: 10.1007/s00122-023-04479-8
• Publication Date: 2023-10-25T10:02:19Z
• Authors: Hermann Gregor Da...; Franziska Löschen...; Herbert Bistrich; Christian Ametz; Herbert Hetzendorfer; Laura Morales; Sebastian Michel; Hermann Buerstmayr
• Keywords: Genomic Selection; Phenome; Phenomics; Predictive modelling

Abstract Key message NIRS of wheat grains as phenomic predictors for grain yield show inflated prediction ability and are biased toward protein content. Estimating the breeding value individuals using genome-wide marker data (genomic prediction) is currently one most important drivers progress in major crops. Recently, technologies, including remote sensing aerial hyperspectral imaging plant canopies, have made it feasible to predict absence genetic data. This commonly referred prediction. Hyperspectral measurements form near-infrared spectroscopy been used since 1980 s compositional parameters harvest products. Moreover, recent studies from was yield. The same showed that can outperform genomic genome static not environment dependent, thereby limiting ability. Gene expression tissue specific differs under environmental influences, leading a tissue- environment-specific phenome, potentially explaining higher predictive Here, we compare variety traits We predictions some traits. However, information present predictor. Future on this topic should investigate whether population retained they Furthermore, find unbiased abilities considerably lower than previously reported recommend method circumvent issue.