Problems interpreting WaterSwap Results

[Javier García Marín]

Hello to everyone!

I have experimented some troubles analyzing waterswap results and I don't know how interpret them. When I visualise per residues results n the Chimera Plugin I have observed some residues presents really high water-cluster energies ranging from 13.456-1786.56 comparing with ligand favourable residues energies (ranging from -0.1123 to -3.54). I don't really know how interpretate these results because final Bennets, FEP, and TI energy to swap out the ligand are 16.65, 16.85, 17.91. Moreover, some residues are associated with a -NaN energy ( I think it means is really high but Im not sure). I have  used an equilibrated coordinates from molecular dynamics using ff14SB forcefield for protein and GaFF for ligand. 

Thanks for your help

Cheers

Javier 

[Christopher Woods]

Are you talking about residue energies in individual results-x.log files or the average calculated using the plugin (or by hand). Individual results.log energies vary widely because of statistical error. These get smoothed down by averaging across lots of results-x.log files.

  Best wishes,

  Christopher 

--
You received this message because you are subscribed to the Google Groups "Sire Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sire-users+...@googlegroups.com.
To post to this group, send email to sire-...@googlegroups.com.
Visit this group at https://groups.google.com/group/sire-users.
For more options, visit https://groups.google.com/d/optout.

--

Sent from Gmail Mobile

[Saad Raza]

Hi,

I checked the values with Javier, by taking average from plugin some residue values are quiet large.

Regards

Saad Raza

[Christopher Woods]

  Hi Saad,

Taking a look, the PMF looks a little discontinuous (large changes near lambda=0 and lambda=1). It looks from the energies that some of the swapped water molecules may be overlapping with the residues in the protein. Have you looked at the PDB structures and the swapped water molecules? How close are they to the residues that have large free energy components? You could experiment with reducing the number of swapped waters by manually specifying the identity atoms (using the "identity atoms" option - see waterswap --help-config). You just pass in the names of the atoms that should have attached identity points (and hence water molecules).

  Best wishes,

  Christopher

p.s. Apologies for the slow reply - at the CCPBioSim conference

---------------------------------------------------------
Christopher Woods
+44 (0) 7786 264562
http://chryswoods.com

[Javier García Marín]

Dear Christopher, As you suggested me, I have checked PDB files with Saad and there is some SWP molecules overlaping with the ligand, to manually specifiying indentity atoms waht is the appropiate format for the config file?

Thanks for your help Christopher

Cheers 

Javier 

[Christopher Woods]

  Dear Saad,

The swap water molecules are supposed to overlap with the ligand. This is how the method works (it swaps the ligand with an equivalent volume of water that occupies the same volume). The issue is that the swap water molecules look like they are overlapping with the protein.

I've looked at the bound_mobile_001500_0.00500.pdb file and the swap water cluster does look weird. This is because you have a long, thin ligand. The swap water algorithm performs poorly for long, thin ligands. Ideally the swapped water should form a dense cluster that fills the same volume as the ligand. In your case, the swapped water molecule has spread out and is not sitting underneath the ligand. You can see this in detail by comparing a VDW representation of the swapped water molecules with a VDW representation of the ligand. Ideally they should overlap well and all of the swapped water molecules should be hydrogen bonding to each other. In your case, the swapped water molecules have broken up into three clusters, with too few on the left and bottom in the attached picture, and too many water molecules in the top right (which forces the water molecules to be too densely packed, and hence overlapping with the protein).

This has happened because the automatic method of placing identity atoms is very simplistic and doesn't work well for long thin ligands. You will need to decide on how many swapped water molecules to use manually and place them manually. You place water molecules onto atoms using this syntax in the config file;

identity atoms = ['N2', 'C4', 'C24', 'C20', 'C18', 'C16', 'C13', 'C10', 'C9', 'C6', 'C1', 'N6', 'N5', 'N4', 'N1']

If the config file is called "config" then you can pass it to waterswap using "-C config" on the command line. You should use the names of the atoms in the ligand on which you want to put identity points, so the above config line says to swap the ligand with 15 water molecules, using atoms N2, C4 etc. etc.

  Best wishes,

  Christopher