Identifying Conserved and Variable Residue Positions by Anchor-Centered Mapping of Enzyme Structures

The rapid expansion of structural databases and increasing availability of predicted protein structures create a growing need for systematic and scalable comparative analysis. Here, we report a computational algorithm that identifies recurrent spatial positions and distinguishes geometrically conserved from chemically conserved and variable residue positions. Importantly, the approach relies on geometric comparisons without requiring conventional multiple sequence alignment. Applied to diverse enzyme families and both experimental and predicted structures, the framework enables systematic analysis of protein environments beyond the immediate active site and facilitates the identification of candidate positions for rational protein engineering.