Predicting Behaviour

Designing functional protein systems remains a major challenge due to the difficulty of predicting how proteins will behave in complex biological environments. Interactions at biomolecular interfaces are highly sensitive to local structure, physicochemical properties, and environmental context, making rational design inherently difficult. As a result, many approaches to protein modification and formulation remain empirical, limiting efficiency, reproducibility, and translation.

Our work addresses this challenge by developing predictive frameworks that link molecular structure to interfacial behaviour. Building on mechanistic insights into how protein structure governs reactivity and interaction, we integrate physicochemical descriptors, structural information, and experimental data to model and anticipate protein behaviour at interfaces.

This includes the development of PRELYM, an open-source computational platform that enables prediction of modification outcomes and guides the rational design of biomolecular systems. By capturing key features that govern interfacial interactions, PRELYM provides a route to predict how proteins will respond to chemical modification and formulation strategies, reducing reliance on trial-and-error approaches.