On the usefulness of genomic selection for rice ratoon performance in early breeding stages

Abstract Rice ( Oryza sativa L.) ratooning, a sustainable production system involves regrowing a second rice crop and it is a very common practice in southwest United States. Employing modern tools such as genomic selection (GS) can enhance breeding efficiency by enabling early selection. The Louisiana State University Rice Breeding Program has traditionally focused on developing superior varieties for the Louisiana's rice industry, however ratoon (RT) performance has typically been considered only in the late breeding stages, when there is little genetic variability available, and all the previous selections were made based on other qualitative and quantitative traits. Therefore, we aimed to verify if our pipeline for variety development is efficient in simultaneously selecting top grain yield performance lines for both harvest seasons: the main crop (H1) and the RT. In this context, we tested the following approaches: 1) Selection index, 2) Indirect selection, and 3) GS. Grain yield data evaluated over three years and three locations from the MP6‐8 population was used in this study. The results highlighted the genetic potential to be explored and the reliability of the data quality. Despite the low phenotypic and genotypic correlations between the first and second harvests (0.11 and 0.12, respectively), the plant response indices proved inefficient for dual‐season selection. Consequently, genotype ranking changed between harvest seasons, suggesting their relative independence. In simpler terms, the genotype that yields the highest productivity for H1 may not necessarily be the same for RT. Our study highlights the feasibility of using GS tools to perform early selections for RT and underscores it as a target trait in the breeding decision‐making process.