You are here

Unsatisfied interface H-bond in Ligand docking

2 posts / 0 new
Last post
Unsatisfied interface H-bond in Ligand docking
#1

I am leaning docking with 3dau. I followed the documentation: I went to docking after repacking the cofactor using ligand_rpkimin and then appended the ligand cofactors.

~/rosetta-3.4/rosetta_source/bin/ligand_dock.linuxgccrelease -database ~/rosetta-3.4/rosetta_database/ -s 3dau_aligned_plus_nap_plus_mtx.pdb -in:file:extra_res_fa ~/Desktop/docking_3DAU/NAP.params ~/Desktop/docking_3DAU/MTX.params -ex1 -ex2 -ex1aro -nstruct 1 > out.log

I got something interesting in the log; But there were NO errors or warning:-

protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: ASP 27 OD1
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: ILE 94 O
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: NAP 160 O14
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: MTX 161 N6
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: MTX 161 N1
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: MTX 161 N2
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: MTX 161 N3
protocols.ligand_docking.LigandDockProtocol: Unsatisfied interface H-bond: MTX 161 N4

What is it saying and should i be worried about this?

Post Situation: 
Thu, 2013-07-18 12:04
Prasanth Kumar

It's simply an informational message enumerating what potential hydrogen bonding atoms are not hydrogen bonded in that docked conformation.

Generally speaking, in the apo state both the ligand and protein have all potential hydrogen bonding atoms hydrogen bonded (to water if not something else). When you transition into the bound conformation, they still need to be hydrogen bonded, as the loss of a hydrogen bond results in a significant energy penalty. This can either be with water still, or across the interface.

As Rosetta uses implicit solvation, it can enumerate all of the ligand-protein and protein-protein hydrogen bonds, but not the ligand-water and protein-water hydrogen bonds. This means that things that should be hydrogen bonded to water may not get the correct penalty if the protein-ligand geometry would preclude water positioning at that location. To help tease out these issues, the ligand docking protocol uses a surface-area based heuristic to identify those hydrogen bondable atoms that 1) aren't making hydrogen bonds to protein or ligand residues and 2) aren't clearly in a position to hydrogen bond to water.

The total number of buried unsatisfied polars are listed in a column of the output silent/scorefile, and you can evaluate the various docked conformations based on how they compare. (Typically this won't be exactly zero, as the heuristic isn't necessarily perfect.) The tracer output lines you show are just an enumeration of that the unsatisfied bonds are for that particular docked pose, in case you want to examine those atoms in more detail.

My guess is that those lines are from a relatively bad docked conformation. The way to use those lines is to find a particular output structure that you're interested in (from the thousands that you've generated), and then look through the log for the lines associated with that particular structure. I would ignore those lines for all of the thousands of structures which you've already discarded as bad, as they don't tell you anything extra.

Thu, 2013-07-18 14:20
rmoretti