Home
This Title All WIREs
WIREs RSS Feed
How to cite this WIREs title:
WIREs Comput Mol Sci
Impact Factor: 14.016

Water models for biomolecular simulations

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

Modern simulation and modeling approaches to investigation of biomolecular structure and function rely heavily on a variety of methods—water models—to approximate the influence of solvent. We give a brief overview of several distinct classes of available water models, with the emphasis on the conceptual basis at each level of approximation. The main focus is on classes of models most widely used in atomistic simulations, including popular implicit and explicit solvent models. Among the latter, nonpolarizable N‐point models are covered in most detail, including some recent methodological advances and nuances. Notes on practical availability and usage in biomolecular simulations are included. Atomistic simulations that were hardly possible only a short while ago have revealed significant problems that can be traced to deficiencies of most commonly used N‐point water models. Recently developed models of this class approximate experimental properties of liquid water much closer than before, and show promise in practical biomolecular simulations. Obstacles to wider adoption of these more accurate water models, both technical and conceptual, are discussed. It is argued that verifying robustness of simulation results to the choice of water model can be of immediate benefit even in the absence of a clear replacement for older models; a specific strategy is proposed. The review is concluded with a discussion on how force‐field development efforts can benefit from better solvent models, and vice versa. WIREs Comput Mol Sci 2018, 8:e1347. doi: 10.1002/wcms.1347

This article is categorized under:

  • Structure and Mechanism > Computational Biochemistry and Biophysics
  • Molecular and Statistical Mechanics > Molecular Dynamics and Monte‐Carlo Methods
Two principle approaches to representing aqueous solvation in biomolecular simulations. Left: Explicit solvation, in which the biomolecule of interest is embedded in a large ‘box’ of discrete solute molecules, and Right: Implicit solvation, which treats solvent as a structureless continuum with the dielectric and nonpolar properties of water.
[ Normal View | Magnified View ]
Two basic way to add polarization to a nonpolarizable ‘base’ water model. Left: Single Drude oscillator model. A charged particle (blue sphere, charge Q D) is attached to the oxygen site (with adjusted charge −2qQ D) by a stiff spring, creating an effective dipole that responds to the polarizing electric field E. Right: A polarizable dipole model, in which inducible dipole(s) are added directly to the fixed charge center(s), their values adjust self‐consistently in response to the net electric field.
[ Normal View | Magnified View ]
The quality score distribution of test 3‐point rigid water models in the space of dipole (μ) and quadrupole (Q T) moments. Each fine grain point on the plot represents a 3‐point model tested. Scores (from 0 to 10) are calculated based on the accuracy of predicted values for six key properties of liquid water. The resulting proposed optimal model is termed OPC3. For reference, the μ and Q T values of commonly used and the very recently developed 3‐point water models (triangles, quality score given by the color at the symbol position) are placed on the same map.
[ Normal View | Magnified View ]
Calculated temperature dependence of liquid water properties compared to experiment for several rigid water models. (a) Bulk density. (b) Static dielectric constant. (c) Heat of vaporization. (d) Self‐diffusion constant. TIP4P‐Ew results are from Ref TIP5P from Refs TIP3P from Refs SPC/E from Refs and OPC from Ref
[ Normal View | Magnified View ]
The quality score distribution of test water models in the space of dipole (μ) and quadrupole (Q T). Each fine grain point on the plot represents a model tested. Scores (from 0 to 10) are calculated based on the accuracy of predicted values for six key properties of liquid water. The resulting proposed optimal model is termed OPC. The actual positions of AIMD1 and TIP5P are slightly modified to fit in the range shown.
[ Normal View | Magnified View ]
Charge distribution of water molecule in the gas phase from a quantum‐mechanics calculation. Counterintuitively, locations of three point charges that optimally reproduce the electrostatic potential of this charge distribution are very different from the on‐nuclei charge placement used by common water models, e.g., Figure .
[ Normal View | Magnified View ]
Relative error to experiment in various properties of several widely used and more recent rigid water models. Values of the errors that are cutoff at the top are given in the boxes.
[ Normal View | Magnified View ]
TIP3P: widely used 3‐point water model. The model geometry, i.e., |OH| distance and HOH angle, closely approximate experimental values for water molecule in gas phase.
[ Normal View | Magnified View ]

Browse by Topic

Molecular and Statistical Mechanics > Molecular Dynamics and Monte-Carlo Methods
Structure and Mechanism > Computational Biochemistry and Biophysics

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts