Part 2: Calculating Electrostatic Fields
Duration: ~30 minutes
This tutorial will generate a 3D grid of points centered on your molecule.
The electostatic potential felt by each grid point
from the partial charges of each atom is then calculated.
Finally, the surface of the protein is created and colored based upon the electrostatic values.
How the electrostatic potential calculation is accomplished:
Consider two points in space that are fairly close together, both of which have a partial charge:
- One point "radiates" it's electrostatic potential, and the other point "feels" this potential. The second point
"reacts" (via dipolar reorientation and electronic polarization) to this potential and
the electrostatic potential of the second point changes in value.
- The second point, with it's new value, "radiates" it's potential, and the first point "feels"
this potential. The first point "reacts" to this potential, and the electrostatic potential of the
first point changes in value.
- Since the first point has changed it's electrostatic potential value,
the calculation returns to step 1 with this new value of the
first point. This calculation loop is repeated until the changes in the electrostatic potential values
are very small.
The total electrostatic potential is equal to the electrostatic charge of the atom plus the electrostatic
potential generated by this "reaction field" (caused by the presence of the other atoms),
as described by the Poisson-Boltzman equation.
More information about this method is available from the DelPhi manual in room A701.
Each grid point is calculated based upon it's
"reaction" to the values of the 6 neighboring grid points. This causes two problems:
- Since we can't calculate
an infinite grid, we must have a boundary to our grid. Points at the boundary of the grid
only have 3, 4, or 5 grid point neighbors, causing inaccuracies. (Which grid points have 3 neighbors?
Which have 4? Which have 5?). To reduce this error, we can make a grid with boundaries that are far from
the molecule, which resides at the center of the grid.
- Since we are essentially extrapolating the electrostatic potential values from one
grid point to another, the grid points should be very close together (i.e., the grid
resolution must be very small) to avoid any "extrapolation" or "round-off" error.
Unfortunately, if we try to avoid both problems, we need to create a very large grid with very many grid
points. This can take a very long time to calculate.
For the purposes of this tutorial, we will calculate one grid with a very large size and
only average resolution. If you attempt this calculation and you want research-grade results, you can
apply the following trick:
- Calculate a very large grid with only average resolution, as described below.
- Then repeat the calculation, select a smaller grid (e.g., a grid that is
not much larger than the size of
the protein), set the Boundary to focussing, and set the focussing grid to the grid that you
calculated in step 1.
This will set the boundary points of the second (focused) grid to values
interpolated from the nearest points found in the first (focussing) grid.
Thus, the boundary points of the second grid will feel an approximation of the
"reaction field" of the grid points of the first (focussing) grid that lie outside the second
(focused) grid. This method has been shown to work very well.
The tutorial performs these steps:
- set up the system
- create a grid
- calculate the electrostatic potential at each grid point
- create a surface of the zinc finger
- color the surface based upon the grid points nearest to the surface.
- If you are not running InsightII from stereo3 or splatter,
see the instructions in the