Correlation and Path Analysis of High-Throughput Phenotyping Traits, Lint Yield and Yield Components in an Exotic Cotton Population

Previous studies have demonstrated that high-throughput phenotyping (HTP) measures, particularly canopy spectral reflectance and temperature, correlate with cotton lint yield. However, meaningful analysis and interpretation of these multi-dimensional and, often, collinear data can be challenging. Thus, this study aimed to 1) quantify the phenotypic diversity of a day-neutral exotic cotton (Gossypium hirsutum L.) population for plant development and agronomic performance under varying soil moisture conditions and 2) investigate path analyses to characterize the relationships between high-throughput phenotyping traits, lint yield, and yield components. Two-hundred and sixteen exotic cottons and 8 cultivars/germplasm lines were evaluated at Jackson Springs, NC in 2016 and 2017 under well-watered and water-limited conditions. Trials were arranged as an alpha lattice design with 3 replicates. In addition to agronomic performance, canopy spectral reflectance and height were measured during the effective flowering period (~45 to 80 days after planting) on a weekly basis using a ground-based HTP platform. Leaf area index (LAI) for each collection date was estimated from the canopy reflectance and height measures; and three developmental traits, including LAI at 45 days after planting, early growth rate (LAI 45 to 65 days after planting), and late growth rate (LAI 65 to 80 days after planting), were calculated based on the LAI estimates. A mixed effect model was used to estimate genotype effects, and path analysis was used to estimate the direct and indirect relationships between the developmental traits and lint yield. Overall, the results demonstrate that early growth rate had a significant, positive indirect relationship with lint yield under both irrigation regimes, primarily mediated by a direct relationship with boll number. Our results also highlight the potential for path analysis as simple and practical method to derive selection indices for agronomic performance based on HTP measures.