Accelerating computations to order N for electrostatic interactions of biomolecular systems

COMP 107

Benzhuo Lu, blu@mccammon.ucsd.edu, Chem. & Biochem, university of California at San Diego, 9500 Gilman Dr, La Jolla, CA 92093-0365, Xiaolin Cheng, xcheng@mccammon.ucsd.edu, Department of Chemistry and Biochemistry, University of California, San Diego, Urey Hall 4206, 9500 Gilman Drive, La Jolla, CA 92093, Jingfang Huang, Department of Mathematics, University of North Carolina, Chapel Hill, NC 27599, and J. Andrew McCammon, jmccammon@ucsd.edu, Department of Chemistry & Biochemistry and Department of Pharmacology, University of California, San Diego, 9500 Gilman Drive -- Dept. 0365, La Jolla, CA 92093-0365.
An order N algorithm both in cpu time and storage is developed and implemented for the calculation of electrostatic interaction of biomolecular system. A boundary integral approach is used to solve the linearized Poisson-Boltzmann equation, in which the new version of fast multipole method (FMM) is applied for far field integration. Proper integral formulas are also given to describe the interacting system with more than one macromolecules, the algebraic system generated by which is well conditioned, and fits well the present FMM frame and can be treated similarly as in the single molecule case. Besides the electrostatic interaction energy, we also show that the force and torque calculations, through computing the surface stress tensor by an interpolation method as a post-processing, are also to O(N). This approach enhances the present computational ability to treat electrostatics of large scale system, such as in protein-protein binding, and hopefully supply a potential way for mocelcular dynamics or Brownian dyanmics simulations of these processes with full PB forces.