Premium
Rapid approximation to molecular surface area via the use of Boolean logic and look‐up tables
Author(s) -
Le Grand Scott M.,
Merz Kenneth M.
Publication year - 1993
Publication title -
journal of computational chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.907
H-Index - 188
eISSN - 1096-987X
pISSN - 0192-8651
DOI - 10.1002/jcc.540140309
Subject(s) - sasa , surface (topology) , atom (system on chip) , series (stratigraphy) , potential energy surface , range (aeronautics) , point (geometry) , computer science , accessible surface area , mathematics , algorithm , computational chemistry , chemistry , parallel computing , molecule , physics , materials science , geometry , quantum mechanics , composite material , biology , paleontology
Abstract We report the development of a new approximate method of calculating molecular surface areas. Our technique is based upon the method of Sharake and Rupley but incorporates several major advances. First, we represent the state of surface points as bits in a bit string so we can utilize Boolean operations to simultaneously turn off multiple test points in one Boolean AND operation. Second, we use a series of Boolean mask look‐up tables to reduce the time complexity of the calculation of molecular surface area down to the same magnitude as doing a potential energy evaluation. When we use a 256 surface point sphere for all of the atoms in BPTI, a 454 nonhydrogen atom protein, and a 1.4‐Å solvent probe, we in general underestimate the total solvent‐accessible surface area (SASA) by approximately 1.25% with a correlation coefficient of 0.9990 over a wide range of conformations. The average CPU time required to calculate the SASA of a BPTI conformer is 0.58 s on an SGI 4D/220 workstation. We also describe a method by which we can calculate an approximate finite difference SASA gradient for BPTI in 0.79 of CPU time. © 1993 John Wiley & Sons, Inc.