Protein dynamics inform protein structure: An interdisciplinary investigation of protein crystallization propensity

Progress and potentialIn this study, we explicitly resolve protein dynamics to capture the critical determinants of protein crystallization propensity through an interpretable attention-based graph neural network model. We show here that proteins must be considered as dynamic moieties and that this essential attribute plays a pivotal role in resolving their crystallization propensity. This is the first work to use structural dynamics features for crystallization propensity prediction. We introduce DSDCrystal, a new toolbox for protein crystal quality prediction, encoded directly with protein dynamics as key input features. Our predictive tools may enable the rational design of protein sequences that result in a diffraction-quality crystal by considering comprehensive biological mechanisms. This framework expands the classical paradigm of structural biology and establishes a roadmap for layered and intuitive control for functional protein design.Highlights•Framework merges physics and ML to predict crystallization propensity via protein dynamics•An interpretable protein crystallization propensity predictor validated by MD simulation•New insights into how dynamics influence protein structure characterizationSummaryThe classical central paradigm of structural biology links a protein's sequence to its structure and function but overlooks conformational fluctuation that is integral to protein function. We propose a graph neural network model based on gated attention that explicitly incorporates protein dynamics via physics-based models to predict protein crystallization propensity. We compare results to all-atom molecular dynamics simulations of flexible, disordered human tropoelastin and ordered, globular human lysyl oxidase-like protein. Our findings show that fluctuating residues correlate with locally maximal attention scores in the neural network. By methodically truncating the sequences, we establish correlations between dynamical and physicochemical molecular properties and protein crystallization propensity. Accounting for comprehensive biological mechanisms, our tool can facilitate the rational design of protein sequences that lead to diffraction-quality crystals. Our study showcases the integration of physics-based and machine learning models for structure and property prediction, expanding the classical paradigm of structural biology.Graphical abstract