Selection of white oat lines based on agronomic ideotype using high-throughput phenotyping

Abstract

Abstract The need for more efficient and precise genotype selection in plant breeding has increased due to the demand for scalable approaches that reduce time, cost, and labor without compromising accuracy. High-throughput phenotyping using vegetation indices has emerged as a promising alternative, although its effectiveness depends on proper model calibration. The objective of this study was to select white oat genotypes based on the agronomic ideotype obtained using high-throughput phenotyping, aiming to optimize yield components. The experiment was conducted with 398 F6 lines and four control cultivars under an augmented block design, with field traits and spectral data collected using UAV-based RGB imaging. Vegetation indices were extracted from orthomosaics, and Stepwise multiple regression and MGIDI index were applied. The results revealed phenotypic variability and identified indices such as NGRDI, BX, GRAY2, NRBDI, and RGRI as strongly associated with agronomic traits. A total of 36 lines matched the ideotype criteria, with five showing superior performance. These results indicate that high-throughput phenotyping combined with multivariate selection is effective, although model recalibration is required for different conditions.