Chile peppers (C. annuum L.) are valued for their capsaicinoid content, which contributes to their pungency (heat) and has various health benefits, including anti-inflammatory and anti-cancer properties. Assessing photosynthetic efficiency through the LICOR-600 porometer/fluorometer (https://www.licor.com/env/products/LI-600/) provides insights into the physiological vigor of the plants. This study employs a comprehensive suite of machine learning models to investigate the correlation between photosynthetic efficiency (stomatal conductance and chlorophyll a fluorescence) and Scoville Heat Units (SHU) to predict the capsaicinoid content within 20 chile pepper varieties. Photosynthetic data were collected at two sites, Fabian Garcia Science Center and Leyendecker Plant Science Research Center, Las Cruces, NM, with readings taken from three different leaves of each of five plants per genotype. Capsaicinoid levels were quantified using High-Performance Liquid Chromatography (HPLC) for each variety. Correlation and principal component analyses (PCA) were implemented to discern the primary influencers on capsaicinoid production. Five predictive models were explored: Decision trees, Random forests, Ridge regression, LASSO Regression, and Support Vector Regression. Each model was applied to predict both total SHU values and categorical SHU labels (mild, hot, very hot). Among these, the decision tree model was the most superior, achieving an R² of 0.77. Initial findings indicate notable variability in photosynthetic activity and capsaicinoid concentrations across the varieties, suggesting a significant but complex relationship that may guide future genetic improvements. The challenges in modeling can be attributed to data collection constraints. Additionally, uniform growing conditions across all test plants might have limited the variability necessary for more definitive model differentiation. This analysis not only advances our understanding of the physiological and genetic factors affecting capsaicinoid content but also underscores the complexities of modeling agricultural traits under consistent environmental conditions. Future research should consider more frequent data collection and the introduction of environmental stressors to better capture the dynamics influencing capsaicinoid production in chile peppers. Key word: High-Performance Liquid Chromatography, Scoville heat unit, photosynthetic efficiency
