Reliability and Usefulness of Seedling Plant Data in Cassava Breeding.

Cassava breeding traditionally involves several years of phenotypic evaluation and field selection. The process begins with seedling evaluation trials (SETs), followed by single row trials (SRTs), and then progresses through preliminary, advanced and uniform yield trials. In recent years, significant efforts have been made to shorten the cassava breeding cycle through genomic selection, which relies on genotypically estimated breeding values (GEBVs). Breeders have typically performed visual selection during the SET stage, when each genotype in segregating populations is represented by a single plant. Despite the intense selection pressure applied at this stage-often eliminating 80-90% of genotypes-no data are collected prior to selection. As a result, breeders miss the opportunity to assess the degree and direction of dominance for key cassava traits, which remains largely unknown. This study provides pioneering insights based on seedling data collected over three years of field evaluations, along with the performance of selected genotypes at the SRT stage. Beyond its relevance for conventional inheritance studies, SET data can also be used to estimate phenotypically estimated breeding values (PEBVs) of progenitors, serving a similar purpose to GEBVs in genomic selection strategies. In fact, these two approaches to estimating breeding values can be highly complementary.