[Chimera-users] inter-surface calculations in Chimera

[Elaine Meng]

Hi Gali,
Eric forwarded your question to me.  The Intersurf tool generates a  
surface between the interacting molecules that is different from the  
surface of either molecule - it is mainly useful for viewing an  
interface and showing its properties (for example, with coloring by  
hydrophobicity or electrostatic potential), but not for measuring the  
surface area.  It is possible to measure volume and surface area in  
Chimera, but currently it is a long process. We hope to have a simpler  
process available in the future.  I'll describe two different  
approaches.

(1) the simpler and more approximate way to calculate buried surface  
area and void volume in the complex:

approx buried surface area = [(surface area of X) + (surface area of Y)  
- (surface area of XY)] / 2

approx interface void volume = (volume of XY) - (volume of X) - (volume  
of Y)

The "Measure Volume and Area" tool can be used to calculate the surface  
areas and volumes.
[  
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/measurevol/ 
measurevol.html ]
The trick is to first make a surface that encloses both X and Y and  
measure its values.  Then, make surfaces that separately enclose X and  
Y and measure again.

Here is a concrete example.  PDB structure 1cf7 has two protein chains  
A and B, two DNA chains C and D, and some water.  I'd open the  
structure and delete everything but the protein and DNA.

command:  open 1cf7
command:  select :.a-d
command:  sel invert
command:  delete sel
command:  surfcat complex :.a-d
command:  surf complex
Tools... Surface/Binding Analysis... Measure Volume and Area (measure  
values for the complex)
command:  surfcat protein :.a-b
command:  surf protein    (choose the protein surface in Measure Volume  
and Area, compute again)
command:  surfcat dna :.c-d
command:  surf dna      (choose the dna surface in Measure Volume and  
Area, compute again)

Finally, use the equations above.  All the results can be viewed in the  
Reply Log (under Favorites).  These values are for molecular surfaces  
(also known as solvent-excluded, where the probe touches), not  
solvent-accessible (traced by the probe center).  For this structure, I  
calculate approx 530 square angstroms surface area buried, approx 730  
cubic angstroms of interface void volume.

(2) A more detailed approach can yield different values for the areas  
buried on X and Y:

(surface area of X buried in the XY interface) = (surface area from all  
atoms of X in surface enclosing XY) - (surface area from all atoms of X  
in surface enclosing just X)

with a similar equation for Y.

Currently, the only way to get atomic rather than total surface areas  
in Chimera is by using the "Area/Volume from Web" tool
[  
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/surfvol/ 
surfvol.html ]
which uses a Web server at NCBI and loads the atomic values into  
Chimera.  Because the MSMS option of the server has problems with  
nucleic acids, the only choice for your complexes is to calculate  
solvent-accessible surface areas (not the solvent-excluded surface  
areas calculated by method (1) above).  The main trick is to first send  
only XY to the server (making sure solvent, ions, etc. are NOT  
included).  Then using Attribute Calculator,
[  
http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/calculator/ 
calculator.html ]
sum the values for the atoms of X in that combined surface, and  
likewise get a total for Y.  Then, delete everything except X and send  
it to the server again, getting the total for X alone.  Close/reopen  
the structure, delete everything except Y, and send it to the server to  
get the total for Y alone.  Subtract the value for X (Y) in the complex  
from the total for X (Y) alone.

Here is a concrete example.  PDB structure 1cf7 has two protein chains  
A and B, two DNA chains C and D, and some water.  I'd open the  
structure and delete everything but the protein and DNA.

command:  open 1cf7
command:  select :.a-d
command:  sel invert
command:  delete sel
Tools... Surface/Binding Analysis... Area/Volume from Web
  - calculate "Accessible Surface (Gerstein)" for the "Molecule" which  
now consists of only the protein and DNA
command:  select :.a-b
Tools... Structure Analysis... Attribute Calculator
     calculate an attribute (name doesn't matter) for "models" with the  
Formula:
          sum(atom.accessibleSurface)
     with the option "Restrict formula domain to current selection"
     write down the total for A-B (which is the PROTEIN IN THE COMPLEX)
     (I get approx 7473)
command:  select :.c-d
     calculate an attribute (name doesn't matter) for "models" with the  
Formula:
          sum(atom.accessibleSurface)
     with the option "Restrict formula domain to current selection"
     write down the total for C-D (which is the DNA IN THE COMPLEX)
     (I get approx 4502)
command:  delete :.c-d        now only the protein is left
Tools... Surface/Binding Analysis... Area/Volume from Web
  - calculate "Accessible Surface (Gerstein)" for the "Molecule" which  
now consists of only the protein
  - get the total for PROTEIN ALONE (Favorites... Reply Log)
    (I get approx 8842)
command:  close 0
command:  open 1cf7
command:  select :.c-d
command:  sel invert
command:  delete sel        now only the DNA is left
Tools... Surface/Binding Analysis... Area/Volume from Web
  - calculate "Accessible Surface (Gerstein)" for the "Molecule" which  
now consists of only the DNA
  - get the total for DNA ALONE (Favorites... Reply Log)
    (I get approx 5758)

Use the equation above to get the buried solvent-accessible surface  
areas.  I calculate approx 1369 square angstroms buried for the  
protein, approx 1256 square angstroms buried for the DNA.  You can't  
compare these solvent-accessible areas to the solvent-excluded areas  
calculated by method (1) ... the solvent-accessible values are  
generally much larger.

PHEW!  I congratulate everyone who got this far.  The written  
explanation makes it seem longer than actually performing the process.
I hope this helps,
Elaine

On Nov 30, 2006, at 7:35 AM, galig wrote:

> Dear Eric,
>
> I have a small question regarding the calculation of inter-surfaces in  
> Chimera:
> I have several protein/DNA complexes, and I would like to compare  
> their protein/DNA interfaces. I used Tools-->surface/binding  
> analysis-->compute interface surface and chose the protein and DNA  
> chains. the result, if I understand it correct, is per atom. Is there  
> a place to find the overall number representing the contact surface?  
> is this a different or similar calculation?
> One more thing- how do I calculate the void intermolecular volume  
> between the chains?
>
> Thank you and sorry to bother you,
>
> Gali Golan
> Dept. of Inorganic Chemistry
> The Hebrew U. of Jerusalem
> Jerusalem 91904 Israel
> Tel. 972-2-6585610
> Fax 972-2-6585319
>

-----
Elaine C. Meng, Ph.D.                          meng at cgl.ucsf.edu
UCSF Computer Graphics Lab and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco
                      http://www.cgl.ucsf.edu/home/meng/index.html