Machine Learning Assessment of the Environmental Factors Contributing to Shade Adaptation in Brassica juncea

Brassica juncea is a widely cultivated leafy vegetable species in Northeast Asia, including Korea, Japan, and China. Under shade conditions, B. juncea exhibits shade avoidance syndrome (SAS), which negatively impacts its market quality. However, B. juncea is cultivated in diverse climates worldwide, including regions with frequent foggy days, highlighting the need to understand its adaptation to shade conditions to improve cultivation quality. To investigate the relationship between SAS phenotypes and environmental factors, including daylength, precipitation, and temperature, we analyzed 30 clones and six commercial cultivars of B. juncea. After 7 days of growth, all six commercial cultivars exhibited a canonical SAS response, with hypocotyl length increasing by 3.25- to 5.18-fold under dim light compared to white light conditions. Among the 30 clones, shade responsiveness varied widely, with hypocotyl elongation ranging from 1.42- to 8.54-fold change. A simple correlation analysis revealed that environmental factors were not highly correlated with shade responsiveness due to their complex interactions. To address this, we applied six machine learning models and found that the random forest algorithm provided the most accurate predictions of environmental influences on hypocotyl length. Using this model, we identified daylength, precipitation, and temperature as key environmental factors contributing to SAS phenotypes in B. juncea. Our findings not only identify clones that can be cultivated under low-light conditions with reduced SAS effects but also establish a link between SAS phenotypes and natural environmental conditions. These insights provide a foundation for future breeding strategies to improve shade adaptation in B. juncea.