GWAS and genomic prediction: Breeding strategies for maize stover digestibility

On the European Atlantic coast, maize silage is essential in the diet of dairy cows, although its digestibility is limited by the strength of cell wall fibers, which directly affects milk production. This study presents an innovative advance in the genetic improvement of forage maize by optimizing both the quantity and quality of lignocellulosic biomass through a combination of GWAS and genomic prediction strategies. Eleven QTLs associated with cell wall components were identified, highlighting genomic regions containing genes related to lignin biosynthesis, polysaccharide modification and delignifying enzyme activity. These findings confirm the polygenic nature of lignocellulosic biomass digestibility and provide the basis for integrating genomic selection into breeding programmes. Genomic prediction using GBLUP exhibited moderate predictive ability for the majority of the fiber traits examined. However, this ability was particularly low for digestibility of organic matter (DOM) and acid detergent lignin (ADL), with the latter accounting for 58% of the variability in DOM. Based on the genetic correlations among traits, a multivariate GBLUP model including hemicellulose was fitted, demonstrating a moderate improvement in the prediction of DOM and ADL. This is one of the first studies to use genomic predictions for this purpose, proposing to optimize the models by including more related individuals in the training set and integrating more complete genotypic data, which will maximize the predictive capacity of the prediction models. • Identification of multiple QTLs associated with cell wall composition and forage digestibility, showing the genetic complexity of these traits. • Detection of key genes involved in lignin biosynthesis and polysaccharide modification, offering new targets for genetic improvement of forage maize. • Proposed improvements in genomic models, including the incorporation of more closely related individuals in the training set and the integration of more complete genotypic data, to optimise predictive ability. • Confirmation of the polygenic nature of maize stover digestibility, establishing a solid framework for the implementation of genomic selection in breeding programmes.