International audience
Haiti's diverse soils and microclimates impact sorghum performance across environments, complicating genotype selection. Therefore, choosing stable genotypes and the physiological determinism associated with them is crucial for breeding programs. Conventional phenotyping for stability is time-consuming and costly. Genomic prediction offers a cost-effective approach, aiming to shorten selection cycles and enhance genetic gains. Limited studies exist on genomic prediction and other selection methods, specifically phenomic selection on sorghum. This paper aims to optimize genomic prediction models for predicting sorghum genotype breeding values in new environments and, alternatively, to test the potential of phenomic selection. In this context, optimization involves selecting the best models and selection indexes.