Skip to main content
eScholarship
Open Access Publications from the University of California

The Semi-Explicit Assembly Solvation Approach and Selected Applications

  • Author(s): Kehoe, Charles Ward
  • Advisor(s): Dill, Ken A
  • et al.
Abstract

We present a new solvation approach, capable of accelerating the computations required for explicit solvent modeling by several orders of magnitude. This technology allows researchers to apply the most accurate solvation models available without sacrificing speed. Semi-Explicit Assembly (SEA) runs at nearly the same speed as the fastest solvation models currently in use. We make this possible by distilling out and reproducing the most important physics modeled by explicit solvent. This approach can be applied to any model of any dipolar solvent, even at the quantum level (though we do not demonstrate that here). Since water is the solvent found in almost all biological interactions, we have so far limited our work to aqueous solvation.

SEA is based on a set of precomputations which sample solvation behavior around single atoms with varying properties, and record the response of explicit solvent to all common sets of atomic parameters. The most important component turns out to be water's reproducible, but asymmetric, response to the electric field on the molecular surface. Treating waters as particles, rather than as a dielectric field, also allows us to reproduce important physical details.

The first half of this work presents all the details of the SEA approach, along with several sets of results demonstrating its potential. We show that, given a set of conformations to work with, we can match the accuracy of explicit solvent at a similar speed as the simple Generalized Born solvation model. We present several sets of molecules which we can model at explicit solvent accuracy, but in which standard approaches like gamma-A and Poisson-Boltzmann (PB) miss important details.

The second half of this work presents additional applications. The first is a blind solvation test using multiple forcefields and water models. We achieved the second-highest accuracy to experimental values in the test, but our approach was also one of the fastest used. The second application is a visualization tool, named SurfMap, which allows users to explore several types of solvation results using interactive three-dimensional models. SurfMap demonstrates additional solvation details which SEA catches, but PB does not.

Main Content
Current View