Surface mesh restricted to reference ligand interface limits alternative binding modes

[Molsidom]

Hi, I’ve noticed that SurfDock builds the pocket mesh by first carving all residues within 8 Å of the input ligand, then keeping only the surface triangles within 3 Å of that same ligand. As a result, any ligand that extends into sub-pockets or binds differently simply finds no mesh where it needs and fails to dock as it should. Did I understand correctly what is happening? If yes, is there a way to properly expand the surface to make it less specific for the input ligand (which by the way usually re-docks well)?

Thanks for the great work on SurfDock

[CAODH]

Hi, I’ve noticed that SurfDock builds the pocket mesh by first carving all residues within 8 Å of the input ligand, then keeping only the surface triangles within 3 Å of that same ligand. As a result, any ligand that extends into sub-pockets or binds differently simply finds no mesh where it needs and fails to dock as it should. Did I understand correctly what is happening? If yes, is there a way to properly expand the surface to make it less specific for the input ligand (which by the way usually re-docks well)?

Thanks for the great work on SurfDock

Thank you for your professional insight. Our current approach is to find molecules with similar interactions to the reference during virtual screening, and at the same time, we hope that the new molecules have new skeletons. However, this does not affect the use of this method to docking with other molecules. Currently, it is difficult to help find new sub-pockets.