Heritability estimation and performance prediction of commercial rose traits using pedigree based best linear unbiased prediction (ABLUP)

Rose cultivation is an economic cornerstone in the Andean region of Ecuador. Due to its economic significance, ongoing genetic improvement of roses is crucial for maintaining competitiveness. However, publicly available studies that use models to predict the performance of varieties in breeding programs are scarce in the Andean region, possibly due to the challenges of implementing accurate phenotyping in a practical framework. Therefore, our study aimed to estimate heritability and predict the performance of Ecuadorian rose varieties using best linear unbiased prediction (BLUP), utilizing a phenotypic database from the company Conectiflor. The analysis included 33 unique rose varieties from nine distinct pedigrees. We assessed traits such as head size, petal count, stem length, and color, finding the highest heritability for stem length. Model diagnostics indicated that the model may not be well-suited for predicting color. Based on BLUP values, we recommend several varieties for use in breeding programs. Additionally, we propose a BBCH-based phenotyping protocol, developed in collaboration with workers at Conectiflor, for large-scale breeding settings. This study presents previously undisclosed data on rose breeding programs and provides valuable insights that can inform future breeding strategies and improve the selection of superior varieties.