Classifying Circumnutation in Pea Plants via Supervised Machine Learning

Climbing plants that locate and attach to suitable support typically outperform those that remain prostrate. The search for support involves a unique movement pattern known as circumnutation, where the plant rotates around a central axis during growth. Although there have been numerous studies investigating the mechanics of circumnutation, how this process is controlled during support-searching remains unclear. Machine learning presents a potential solution for addressing this issue and enabling accurate phenotyping. Despite being in its infancy, machine learning has already demonstrated broad applicability in plant science. In this study, we tested whether simulation-based machine learning methods can capture differences in movement patterns nested in actual kinematical data. We compared machine learning classifiers with the aim of generating models that learn to discriminate between circumnutation patterns related to the presence/absence of a support in the environment. Results indicate that there is a difference in the pattern of circumnutation, depending on the presence of a support, that can be learned and classified rather accurately. We also identified distinctive kinematic features at the level of the junction underneath the tendrils that seems to be a superior indicator for discerning the presence/absence of the support by the plant. Overall, our findings suggest that machine learning approaches are powerful tools for understanding plant movement.