makeDiffGrid

PROFEC is typically used to suggest changes to a ligand that might improve its binding to a given receptor. Since the free energy of binding involves both the bound and free states of the ligand, PROFEC can be used to compare the effects of ligand modification (test particle insertion) in two different states. For example, if we want to compare our reference ligand (A) with a modified ligand A', we use the following thermodynamic cycle:

ligand A (protein) ---dG1 (PROFEC)---> ligand A' (protein)
| |
dGbind (A) dGbind (A')
| |
V V
ligand A (water) ---dG2 (PROFEC)---> ligand A' (water)

We want to know how the modification A -> A' affects the binding free energies. That is, we wish to calculate ddGbind = dGbind (A) - dGbind (A'). By the thermodynamic cycle, we know that ddGbind = dG1 - dG2, so we can use PROFEC to calculate the value of interest, ddGbind.

This is done by generating grids from two different trajectories; first, grids are calculated for the protein-ligand complex. Second, grids are generated for a simulation of the free ligand (in water). Both of these grids must be defined relative to the _same_ set of atoms on the ligand in each trajectory.

Finally, the program makeDiffGrid can be used to generate a third grid file that is the difference of two input grids. For calculating the free energy of binding, one would use

makeDiffGrid
makeDiffGrid

In the grid, negative values indicate grid positions where adding the test particle favors the protein-ligand complex at the expense of the water-ligand interaction. These are the locations where the ligand might be modified to increase the binding free energy. The file, however, gives a more qualitative indication of how charge should be distributed to favor binding; negatively charged groups should be added in regions of positive electrostatic potential, and vice versa.

For other applications of makeGrid and makeDiffGrid, please consult the Radmer and Kollman paper cited above.