[Dock-fans] Questions regarding chemical matching and DOCK3.5 scoring

BA S bas007ster at gmail.com
Thu Sep 6 14:25:44 PDT 2007


I am new DOCK user and I am trying to dock the ligand (colchicine)
from 1sa0.pdb back into the the receptor (sa0) with an RMSD<2. I have
two questions, one is related to chemical matching and the other is
related to the DOCk3.5 score and grid generation.

Here is what i have tried so far.
1)Ligand preparation: As descibed in the tutorial with Chimera.
2)Receptor preparation: Without minimization in Chimera.
1.Removed ligand atoms and unwanted chains (c,d and e)
2.Added hydrogens and charges as described by the tutorial and saved
sa0_charged.mol2 and removed the hydrogens and saved sa0_noH.pdb
3.Molecular surface was calculated and spheres were generated with sphgen_cpp.
The density of the points of the molecular surfaces was changed to 2
(-d 2). The spheres generated covered the receptor better compared to
the normal density of 1, but did not cover certain areas occupied by
the ligand. Therefore tried to minimize the receptor.
3)Receptor preparation: With minimization in Chimera
1.Added H and charges as above.
2.Minimized the whole protein with no fixed proteins
3.Molecular surface was calculated as above and spheres covered the
recptor better that above
4.Spheres were selected around the original ligand ((8A)
5.Box was generated.
4)Grid generation using GRID
1.Grid were generated with GRID
2."energy_cutoff_distance 5" , "contact_cutoff_distance 6" and a
Lennard-Jones of 6-8 yielded the best results.
1.Grid scoring the original ligand without orienting yielded a high
score were the vdw score dominated the score.
(energy_cutoff_dissatance to 5 yielded the lowest electrostatic
scores (Grid score >200)
2.Primary: Contact scoring, Secondary: Grid
1.With the contact scoring i was able to get an RMSD <2 and the
secondary Grid score still high with the vdw score dominating the
score (Grid score > 150)
3.Primary: Grid scoring (automated)
1.Bump filter was activated and left unchanged
2.I was able to get good scores (Grid score = -28.95) but the RMSD
scores were all > 6
4.Primary: Grid Scoring (manual matching)
1.Spheres where marked critical and in some configurations i was able
to get an RMSD<2 and a smaller but still positive score (Grid score >

Question 1
I would like to do chemical matching because the Colchicine
pharmacophore site has been documented. I am however unsure how to
label the spheres. I would like to label certain spheres as hydrogen
donors, hydrogen acceptors, hydrophobic and the rest as uncritical and
neutral. Attached is my spheres file (sa0_spheres.sph). Is there
anything that i need to change in the chem.defn and chem_match.tbl
files? Can someone please give an example of how i would label the
.sph if spheres nr: 1154, 1554, 1558 were acceptors, 487, 1293, 1301,
1433, 1437 were hydrophobic and 1026, 460, 454 were donors?

Question[s] 2 related to DOCk 3.5 scoring.
1)In order to create the CHEMGRID, in INCHEM, does the receptor pdb
need to be protonated?
2)When the CHEMGRID is created do i need to convert it.? If yes, i
tried to do it with gconv.in, but im unsure what the .crg file is or
how it is created. I decided to rename the protonated sa0.pdb file to
sa0.crg and ran the "grid-convert" command. I received a segmentation
3)In creating the electrostatic grid,  i also needed a .crg file
whereby the receptor and selected spheres are included in 1 file.
4)  I created a combined file in Chimera and saved it as a .pdb and
renamed it to .crg file. I then ran the make_phimap command
succesfully with rec_sph.phi as an output. Is thi the correct way of
doing it?
5)Once i have created the electrostatic grid, (.phi)  how do i
incorporate it in the grid creation or creation?

Thank you

More information about the Dock-fans mailing list