Generative Artificial Intelligence for Efficient Prediction of Protein Redox Potentials (RedoxAI)
PI: Gokcen Kestor and Roberto Gioiosa, Physical and Computational Sciences Directorate
RedoxAI seeks to accelerate the development of new materials and catalysts by rapidly estimating the effects of multiple charges in the embedding protein scaffold on the redox properties. This is done by leveraging fine-tuned LLMs to predict the reduction potential from the arrangement of charges as well as the needed arrangement of charges for desired redox properties.