Article ID: 755 - Last Modified: August 19, 2014
The core constraints do not seem to constrain the core to where I want it. Is there a way to force Glide to put the core in the right place, and then sort out the rest of the molecule?
Yes. Starting with the 2014-3 release, when you use a core constraint, Glide "snaps" the core of each ligand to that of the reference ligand, and then adjusts the rest of the ligand. In the post-docking minimization the core is constrained with a flat-bottomed potential that has very steep walls, which permits a small amount of movement from the reference ligand position. (If you want to use the old core constraint method, set CORE_SNAP=true in the Glide input file for the docking run.)
In earlier releases, the core constraint applies a restraining potential, but the docking process doesn't anchor the core to the reference position and sample only the side chains. However, if there are good poses with the core in the desired position, Glide should find them. If it isn't finding good poses, it's possible that the rigid receptor used by Glide is not appropriate for your ligand. You could try superimposing the ligand core on the reference ligand core in Maestro (with Tools → Superposition), and then running Glide in score-in-place or refinement-only mode. This might give you some indication of why the ligand isn't docking with the constraints. Adding per-residue interactions to this run might give you some further diagnostic clues. If the ligand has clashes with the rigid receptor, you could then try induced fit docking (IFD) to relieve the clash.
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles