Electrostatics shape the structure, stability, and dynamics of charged (bio)matter, yet how heterogeneous, anisotropic charge distributions control protein–protein interactions remains unclear. I will present a versatile multiscale framework that links molecular electrostatics to collective properties by combining a colloid-inspired coarse-grained model with neural-network–assisted optimization trained against experimental SAXS data.
Using monoclonal antibodies as a representative anisotropic system, the approach identifies coarse-grained charge patterns that reproduce measured structure factors and predict rheological behavior over a wide range of concentrations. Feature-attribution analysis further reveals which physical descriptors and spatial arrangements of localized charge patches dominate the emergent solution structure.
The resulting, transferable strategy offers a predictive route to decipher charge-driven interactions in anisotropic biomolecules and, more broadly, heterogeneously charged soft-matter systems, with immediate relevance to protein formulation and functional materials design.