Wouldn't it be better to have the Schrödinger fragment library available in two versions: one version that is prepared with and one prepared without metal-binding states?

There actually isn't a need for this, because of the way the metal and normal state penalties are applied in Glide. When generating metal-binding states, Epik actually stores two (or more) state penalties for each metal-binding ligand:

1) The normal state penalty, which will be quite large for the "Metal Only" ligand states; and
2) the metal state penalty, an atom-level property that indicates what the state penalty should be if that atom is bound to a metal.

When docking these ligands, Glide automatically skips the "Metal Only" states if there is no metal in the grid box. If there is a metal in the grid box, all ligands will be docked. If the ligand binds to the metal, the appropriate atom-level metal state penalty is applied. If the ligand doesn't bind to the metal, the normal state penalty is applied (which will be large for "Metal Only" states).

<> Therefore, there is no need for a second version of the library. The library with metal states can be used whether or not there is a metal in the receptor binding site, with minimal effect on speed (i.e., skipping the metal states) and no effect on the scoring.

Keywords: fragment library, metal binding states, metal only

Back to Search Results