A Functional Data Analysis Approach to RuBisCO Engineering
Mitigating and adapting to climate change requires carbon emission control and effective technologies for drawing greenhouse gases from the atmosphere. Here we propose an effective strategy for guiding the rubisco engineering problem, which seeks to improve photosynthesis and carbon sequestration in crops by minimizing photorespiration, a major impediment to crop yields. Photorespiration occurs when rubisco oxygenates rather than carboxylates, thus reducing carbohydrate synthesis to eliminate toxic byproducts. Most plants, including most agricultural crops, exhibit a C3 photosynthetic pathway and have not developed adaptive mechanisms to combat rubiscos tradeoff. The span of low activity rubiscos opens the possibility of engineering less productive crops to express higher activity enzymes. The main experimental challenges in bypassing rubiscos biochemical limitations include its molecular size and complexity, lack of empirical data, and costs of acquiring new data, but recent advances in machine learning and data analytics could allow us to overcome these challenges. In particular, we propose a novel computational approach to inform experimental research into rubisco engineering by employing recently developed techniques from functional data analysis. We show that the separation between high and low activity enzymes can be modeled within the sequence space of rubiscos primary structure, and we further discuss how our approach can guide deeper investigation into the rubisco engineering problem. Future empirical success would simultaneously address major global issues such as rising atmospheric CO2 levels and food insecurity.