Einstein
Michael Holst
Main: Group
Publications
Software
Teaching
Affiliations
Contact
Bio and CV
Related: FETK
CCoM
CSME
Wikis: FETK Wiki
CCoM Wiki
CSME Wiki
UCSD
Search: Google
G-Scholar
arXiv
MathSciNet
UCSD: UCSD
Mathematics
CCoM
CSME
CTBP
NBCR
SDSC
InfoPath
Blink
Libraries
Library
Weather/Sky: CNN Weather
Wayhoo GPS
NOAA GPS
Several potential isosurfaces of an SOD enzyme.

The pictures below are various visualizations of an SOD molecule, and the electrostatic potential fields generated by SOD in ionic solutions. The potential fields were obtained with the help of the Debye-Huckle continuum electrostatic model, which gives rise to the nonlinear Poisson-Boltzmann equation. X-Ray crystallography SOD data from the Brookhaven Protein Database was used as model input, and the nonlinear Poisson-Boltzmann equation was solved numerically, using a global inexact-Newton-multilevel method developed by M. Holst in Applied Mathematics and CRPC at Caltech. Linearized and nonlinear Poisson-Boltzmann models yield substantially different electrostatic potential values near the binding sites, leading to correspondingly different reaction rates predicted by Brownian dynamics simulations. This implies that the full nonlinear model is important for certain modeling situations.

A parallel CC++ implementation of the global inexact-Newton-multilevel method has recently been developed by M. Holst and John Garnett of the Computational Biology Group at the Caltech Beckman Institute. Additional collaborators on this project were Carl Kesselman of the Beckman Institute and Dan Meiron in Applied Mathematics and CRPC at Caltech. Several of the plots below were produced by James Patton, also in the Computational Biology Group at Caltech.


Valid XHTML 1.0 Transitional Valid CSS!