Leveraging leaf spectroscopy to identify drought-tolerant soybean cultivars

Understanding cultivars' physiological traits variations under abiotic stresses is critical to improve phenotyping and selections of resistant crop varieties. Traditional methods of accessing physiological traits in plants are costly and time consuming, which prevents their use in breeding programs. Spectroscopy data and statistical approaches such as partial least square regression could be applied to rapidly collect and predict several physiological parameters at leaf-level, allowing phenotyping several genotypes in a high-throughput manner. We collected spectroscopy data of twenty soybean cultivars planted under well-watered and drought conditions during the reproductive phase. At 20 days after drought was imposed, we measured leaf pigments content (chlorophyll a and b, and carotenoids), specific leaf area, electrons transfer rate, and photosynthetic active radiation. At 28 days after drought imposition, we measured leaf pigments content, specific leaf area, relative water content, and leaf temperature. Partial least square regression models accurately predicted leaf pigments content, specific leaf area, and leaf temperature (cross-validation R2 ranging from 0.56 to 0.84). Discriminant analysis using 54 wavelengths was able to select the best-performance cultivars regarding all evaluated physiological traits. We showed the great potential of using spectroscopy as a feasible, non-destructive, and accurate method to estimate physiological traits and screening of superior genotypes.