Method for the Poisson-boltzmann Equation

Period of Performance: 04/01/1998 - 06/30/1999

Unknown

Phase 1 SBIR

Recipient Firm

Continuum Dynamics, Inc.
34 Lexington Avenue Array
Ewing, NJ 08618
Principal Investigator

Abstract

A fast parallel boundary element (BE) method is proposed for the efficient and robust numerical solution of the linearized Poisson- Boltzmann equation for application to biomolecular electrostatic problems where ionic strength effects are important. In such an analysis, all N boundary elements representing the dielectric boundary at the biomolecular surface interact with each other and thus comprise a many-body problem with 0(N2) storage and 0(N3) CPU costs in conventional solution methods. Both costs are drastically reduced to 0(N) by combining fast multipole algorithms (FMAs) with an efficient iterative matrix-free solution procedure. To date all fast BE analysis have been limited to the solution of the Poisson equation (zero ionic strength). Thus, a key technical advance proposed here is to employ a FMA to calculate the Debye-Huckel interactions between surface elements over the entire range of salt concentration. To make full use of available computational resources, the inherent parallelisum of the fast BE analysis will be exploited and the software oriented for operation upon distributed systems. Results will be presented for both analytically tractable model problems and also for representative biomolecular systems to verify the anticipated computational performance and accuracy. PROPOSED COMMERCIAL APPLICATION The successful completion of the SBIR effort, will make available to government, academia and industry an advanced serial and parallel fast BE software which will be used in electrostatic problems of biomolecular assemblies immersed in ionic solution. The knowledge gained from using such powerful computational tools in electrostatic problems hold the promises of revealing the effects of electrostatic interactions on biological processes and opening new venues for structure-based drug design.