Figure 2.

A schematic representation of the UF blueberry breeding program, integrating phenotypic selection and genomic prediction. The breeding process is conventionally organized in two integrated steps: population improvement and product development. A breeding cycle starts with crosses between outstanding parental genotypes. After that, several stages (I–IV) are required to evaluate the genotype performance. At Stage I, we will use marker-assisted selection targeting traits with simple genetic architecture. Genomic selection will be implemented in Stage II when GEBVs are computed. In advanced selections (Stages III and IV), high-quality phenotyping will be performed to leverage the calibration of genomic prediction models. At these stages, metabolomics and sensory panel analyses will also play an important role in flavor-assisted selections. In the end, elite materials are registered as clonally propagated cultivars. In addition, to shortening the time for product development, GS can be applied to move top-ranked plants directly from Stages II to IV, skipping at least 3 years of evaluation at Stage III. For population improvement, GS can assist in more accurate parental selection at early stages.