Fourier Transform Method for Surface Area Burial

We develop a method that uses the Fast Fourier Transform (FFT) to efficiently calculate changes in the solvent accessible surface area (SASA) at many relative translations of two binding partners. For each binding partner, the buried surface area (BSA) is computed using the cross-correlation of a grid-based discretization of its SASA and of a solvent exclusion grid from its binding partner. This approach improves the tractability of exhaustive interface burial calculations by reducing computational complexity, making it particularly valuable for analyzing large biomolecular systems. We implemented an optimized library for Shrake-Rupley SASA by modifying sasa.cpp from MDTraj. Our approach reduces the cost of evaluating BSA at all translations within a rotation from to . One FFT pass costs 0.41–0.68 µs per translation against 0.14–1.70 s per pose for the direct calculation, a speedup of to that appears to grow with system size. Agreement with the direct calculation is close, with a root mean square error of 3.0 Å at 0.5 Å grid resolution. Possible applications of the method include virtual screening, models of biomacromolecular assemblies and aggregates, studies of protein diffusion, and thermodynamic analysis of large conformational ensembles.