Le 31/01/2011 20:16, Elaine Meng a écrit :

In that paper, they only looked at Coulombic ESP (and its gradient), but one might expect as least as good results with a more sophisticated ESP. However, you (Sumitro) already said that you were using Coulombic rather than the other methods to calculate electrostatic potential.

You can use the "Compute grid" option of Coulombic Surface Coloring to create a grid file, and then do the gradient coloring that Tom describes with the Surface Color tool. http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/coulombic/coulombic.html http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/surfcolor/surfcolor.html

However, that still leaves open the question of whether any ESP gradient is an adequate measure of hydrophobicity.

That J Med Chem 1992 paper is frustrating because it has several graphs where the lines all look the same, so it is hard to tell what is what. Maybe it exists somewhere in color, where you can tell the lines apart, but I only found black and white. Elaine

On Jan 31, 2011, at 11:02 AM, Tom Goddard wrote:

...

Hi Elaine,

Occurs to me that if the electric field strength, ie the gradient of the electrostatic potential, is a useful measure similar to hydrophobicity, that can be viewed currently in Chimera with the surface color dialog using color by "volume data gradient norm". You would need the electrostatic potential calculated for instance by APBS or DelPhi.

Tom

...

Hi Sumitro, I did a little more digging for related information and thought I would share -- however, none of these methods are available in Chimera currently.

... If I understood it correctly, they looked at how successfully Coulombic electrostatic potential and the gradient in that potential can be used to identify hydrophobic areas of surface (areas that bind hydrophobic atoms in the ligand). Both were somewhat successful, but even better was a simple definition of hydrophobic surface as the part of the surface that is not close to hydrogen-bonding groups. This paper is fairly old and the data set of ligand/receptor structures small.

Elaine

Thank you Elaine and Tom for this interesting discussion. Sorry for my misunderstanding of Elaine's first answer to Sumitro.

The work I mentioned (by Goodford) uses a different approach that is not related (at least not directly as far as I can understand it) with the electrostatic potential. They use a "dry" (originally a methane, I think), hydrophobic probe that is rolled over the protein surface and the energy of interaction of this probe with the protein is calculated at every point of a certain grid. From the three main components of the GRID force field only the Lennard-Jones seems relevant here (the other two are electrostatic - but the "dry" probe is neutral - and hydrogen bond potential)

I am not sure the results would be the same as looking at the electrostatic potential, as Tom suggests, and I cannot check it, obviously. Thank you, Elaine, for these references! The notion of hydrophobicity as a measure of distance to hydrogen-bonding atoms seems also interesting (and looks as easier to implement than the GRID approach)

Best regards,

Hi Elaine and Tom,

Thanks for going the extra mile to come up with such a comprehensive reply. This is a lot of info to digest and softwares to explore. Will try them out. Thanks! =]

Cheers, Sumitro

-----Original Message----- From: Elaine Meng [mailto:meng@cgl.ucsf.edu] Sent: Tuesday, February 01, 2011 11:29 AM To: Sumitro Harjanto Cc: UCSF Chimera Mailing List Subject: Re: [Chimera-users] surface hydrophobicity display

One more: the program "hint" (Hydropathic INTeractions). It's embarrassing that I forgot to mention this, since I've actually used the program for DOCK scoring and published the results, which were pretty good. It is not free, however.

Here is the hint website: http://www.edusoft-lc.com/hint/

It comes as a module to other packages such as SYBYL, or as a standalone. I don't know if the standalone includes the option to output 3D maps. Elaine

On Jan 31, 2011, at 9:55 AM, Elaine Meng wrote:

Hi Sumitro, I did a little more digging for related information and thought I

would share -- however, none of these methods are available in Chimera currently.

Miguel Ortiz Lombardia has pointed out that the GRID program

calculates a hydrophobic potential, and sent us the following links:

http://biop.ox.ac.uk/www/lj2000/noble/noble_02.html http://www.moldiscovery.com/docs/vsplus/grid.html

-- the following web server sounded promising, but apparently no

longer exists:

PLATINUM: a web tool for analysis of hydrophobic/hydrophilic

organization of biomolecular complexes.

Pyrkov TV, Chugunov AO, Krylov NA, Nolde DE, Efremov RG. Bioinformatics. 2009 May 1;25(9):1201-2. http://www.ncbi.nlm.nih.gov/pubmed/19244385

-- the following paper discusses what surface properties of the

receptor correlate well with the types of nearby ligand atoms. If I understood it correctly, they looked at how successfully Coulombic electrostatic potential and the gradient in that potential can be used to identify hydrophobic areas of surface (areas that bind hydrophobic atoms in the ligand). Both were somewhat successful, but even better was a simple definition of hydrophobic surface as the part of the surface that is not close to hydrogen-bonding groups. This paper is fairly old and the data set of ligand/receptor structures small.

Definition and display of steric, hydrophobic, and hydrogen-bonding

properties of ligand binding sites in proteins using Lee and Richards accessible surface: validation of a high-resolution graphical tool for drug design.

Bohacek RS, McMartin C. J Med Chem. 1992 May 15;35(10):1671-84. http://www.ncbi.nlm.nih.gov/pubmed/1588550

-- then there are some other complicated functions based on molecular

fragments and distances among them and to the molecular surface:

A new approach to analysis and display of local

lipophilicity/hydrophilicity mapped on molecular surfaces.

Heiden W, Moeckel G, Brickmann J. J Comput Aided Mol Des. 1993 Oct;7(5):503-14.

Estimating and representing hydrophobicity potential Fauchere et al. Journal of Molecular Graphics Volume 6, Issue 4, December 1988, Pages 203-206

If I understood correctly, these have many parameters and the

functional form is not simple. It would not be straightforward to put them into Chimera. Finally, I also saw that the program MDL Chime has something called hydrophobic potential display, but I couldn't find a description of how it really works -- it might just be coloring whole amino acids by hydrophobicity value like Chimera does now, or atoms based on element.

Elaine

Elaine C. Meng, Ph.D. UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco

---

Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users

Institute for Infocomm Research disclaimer: "This email is confidential and may be privileged. If you are not the intended recipient, please delete it and notify us immediately. Please do not copy or use it for any purpose, or disclose its contents to any other person. Thank you."

On Feb 1, 2011, at 10:20 AM, Elaine Meng wrote:

...

On Jan 31, 2011, at 9:55 AM, Elaine Meng wrote:

...

-- the following web server sounded promising, but apparently no longer exists:

PLATINUM: a web tool for analysis of hydrophobic/hydrophilic organization of biomolecular complexes. Pyrkov TV, Chugunov AO, Krylov NA, Nolde DE, Efremov RG. Bioinformatics. 2009 May 1;25(9):1201-2. http://www.ncbi.nlm.nih.gov/pubmed/19244385

Today the PLATINUM web server exists! I can't quite get what I want for viewing in Chimera, however:

Their user manual explains how to download data for display in other programs, including PDBs with the atomic hydrophobicity constants in the B-factor column. I can open those in Chimera and use Render by Attribute to show the values. The constants are only the starting data, however, not the potential...

According to the PLATINUM paper:

Also, additional output is provided for more detailed analysis of selected ligands/ligand poses which will be discussed below. To perform subsequent analysis, the MHP data for each ligand can be downloaded in one of the following formats.

Simple text file where atoms are annotated according to the MHP atom type parameterization.

The pdb file where either atomic hydrophobicity constants or surface MHP values are written to the B-factor column.

Ligand molecular surface represented as a set of dots in pdb or InsightII (Molecular Simulations Inc., 2000) formats.

Grid hydrophobic/hydrophilic potential in InsightII or MolMol (Koradi et al., 1996) formats.

It would seem that from the second part of #2 that you could get the surface MHP values into the B-factor column (and then use Render By Attribute...) though some perfunctory poking around the PLATINUM manual didn't reveal to me how to do that...

--Eric

Unfortunately I can't get a potential map of the receptor that can be used in Chimera. Platinum is quite ligand- or docked-molecule- centric. You can get molecular surface dots in PDB format with the potential as the "bfactor" but it is only for the surface of the ligand, not the receptor. The potential can also be output as a 3D grid, but in some formats Chimera does not read (InsightII or MolMol), and I suspect the grid might be only in the location of the ligand, not enclosing the whole protein receptor. Elaine

---

Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users

Hello Sumitro et al., Dr. Efremov, whose group produces the PLATINUM server, pointed out that you can get a dot surface of the protein by uploading it as the "ligand" and not uploading any "receptor."

PLATINUM server: http://model.nmr.ru/platinum/

When I tried this with my example protein, it worked fine if I set the surface dot density to "low" or "very low" *before* launching the calculation. Using the default "medium" gave an incomplete surface even though the file was extremely large -- probably some memory limit was reached. (To prepare the example input with Chimera, I opened 2gbp, deleted ligand and waters, added hydrogens, saved PDB file). The server said 2 ligand files were uploaded, but I just ignored that.

After the PLATINUM calculation is done, click the floppy disk icon for the "ligand" that is really your protein, save as "molecular surface" of type "PDB surface." It may several seconds to respond or write the file because even at low density, there are lots of dots.

That gives a file in PDB format of dots as "atoms" with hydrophobicity potential in the bfactor column, for coloring with Render by Attribute in Chimera, or with the command "rangecolor".

For display in Chimera it would still be better to get the whole 3D grid of potential values -- grid data could be used to color a solid surface instead of just dots, or to show isopotential contours. Dr. Efremov kindly said that they will take a look at maybe adding another grid format. For now, it is possible to show the protein hydrophobicity potential by coloring its own dot surface as described above.

Elaine

On Feb 1, 2011, at 11:06 AM, Elaine Meng wrote:

I already did that. It is a dot surface of the ligand only where each dot is a PDB "atom". So you can see dots of the ligand surface colored by the potential from the receptor, but not the surface of the receptor which is what we were more interested in (at least, that's what I was interested in). Elaine

On Feb 1, 2011, at 10:59 AM, Eric Pettersen wrote:

...

According to the PLATINUM paper:

Also, additional output is provided for more detailed analysis of selected ligands/ligand poses which will be discussed below. To perform subsequent analysis, the MHP data for each ligand can be downloaded in one of the following formats.

• Simple text file where atoms are annotated according to the MHP atom type parameterization.

• The pdb file where either atomic hydrophobicity constants or surface MHP values are written to the B-factor column.

• Ligand molecular surface represented as a set of dots in pdb or InsightII (Molecular Simulations Inc., 2000) formats.

• Grid hydrophobic/hydrophilic potential in InsightII or MolMol (Koradi et al., 1996) formats.

It would seem that from the second part of #2 that you could get the surface MHP values into the B-factor column (and then use Render By Attribute...) --Eric

...

PLATINUM: a web tool for analysis of hydrophobic/hydrophilic organization of biomolecular complexes. Pyrkov TV, Chugunov AO, Krylov NA, Nolde DE, Efremov RG. Bioinformatics. 2009 May 1;25(9):1201-2. http://www.ncbi.nlm.nih.gov/pubmed/19244385

Today the PLATINUM web server exists! I can't quite get what I want for viewing in Chimera, however:

Their user manual explains how to download data for display in other programs, including PDBs with the atomic hydrophobicity constants in the B-factor column. I can open those in Chimera and use Render by Attribute to show the values. The constants are only the starting data, however, not the potential... Unfortunately I can't get a potential map of the receptor that can be used in Chimera. Platinum is quite ligand- or docked-molecule-centric. You can get molecular surface dots in PDB format with the potential as the "bfactor" but it is only for the surface of the ligand, not the receptor. The potential can also be output as a 3D grid, but in some formats Chimera does not read (InsightII or MolMol), and I suspect the grid might be only in the location of the ligand, not enclosing the whole protein receptor. Elaine