| Probe With O | Home Page: Richardsons' Laboratory | 10/2000 |
This page describes the software we have developed to permit the examination of steric clashes with the program Probe from inside the model building program O. Graphical display of detailed steric interactions, alongside electron density, can provide a valuable source of additional structural information during crystallographic refitting.
Much of the strength of this method comes from explicitly including hydrogen atoms in the contact surface calculations. Previous versions of these tools included parameter files for including hydrogens within O. Since O was not really designed with hydrogens in mind, the program was not always able to reliably work with them. With this version, we have taken a different approach. With each update, atoms in residues within a given radius around the pick center are saved in a file, hydrogens are added to this temporary file with Reduce, and contacts are calculated on these atoms with Probe.
The process is occasionally a little slow (when multiple interacting OHs are within the selection, for example) but it has several advantages. First, it does not require changes to O data files and parameter sets. Second, it does not clutter the display with H atoms, which most crystallographers prefer not to have to look at.
| THIS MACRO MAKES AND REMOVES THE FOLLOWING FILES ANY FILES OF THE SAME NAME WILL BE REMOVED prbsph.new prbsph.idx prbsphH.pdb prbsph.pdb prbsphdots.odb |
The additional software required to use Probe with O is minimal:
| prbsphereconv | a small UNIX awk script, which converts O data formats, and |
| probe.mac | an O macro, which activates pick center and then controls program calls to the script, Probe, and Reduce. |
Both of these files are in probewithO.tar.Z (along with a copy of this file).
You'll also need to have Reduce and Probe running on your computer (or network) and in your UNIX path. If needed, these programs can be obtained from our website: http://kinemage.biochem.duke.edu.
And, of course, you'll need O, directions for obtaining and use can be found from this website: http://imsb.au.dk/~mok/o/
The script and macro discussed here work with version 7 or 8 of O.
When you select @probe.mac from the menu, and click on an atom, it will dump the current coordinates of a 6Å sphere of atoms to a temporary file. It will then run Reduce and Probe to create a graphics file of Hbonds and steric clashes and display the resulting output. Steric clashes are displayed as yellow, red or pink spikes, depending on severity; H-bonds are displayed as green pillows of dots.
Probe has many parameters; for a full listing type: probe -help
The command: probe -oformat -stdbonds -mc -self "all" xxxH.pdb > xxx.odl
will produce a file of O graphics objects (xxx.odl) for van der Waals dots, Hbond dots and steric clash spikes for all the residues in the coordinate file xxxH.pdb (which includes hydrogens added by reduce)
The O command: draw_object xxx.odl
will display these graphics objects
The Probe command line flag, -stdbonds, consults an internal connectivity table for standard residues and bases. If you are working with a portion of your structure which has modified amino acids you will probably want to run Probe without this flag.
Reduce adds hydrogens using simple geometric considerations. Then it identifies groups where the hydrogen position needs to be optimized. When it is trying to add hydrogens to groups like serine and threonine OH groups, Reduce tries to rotate the OH group to best effect. This is a combinatorial process but usually only a few groups interact: we call this a "clique". Usually, Reduce can enumerate and test each possible configuration quickly. In pathological cases, these cliques can be larger than Reduce can handle and it will not optimize the positions of those hydrogens, giving you a warning message instead.
Pathological cases are actually quite rare. If your structure seems to have an enormous clique, check to make sure that separate subunits have different chainIDs (the script pdbcns can be used to convert segids to pdb-format chainIDs). The command line flag
If you would like to examine the hydrogens in your structure, you can add hydrogens to your coordinate file with Reduce using the command:
where xxx.pdb is your PDB format coordinate file. The -build flag causes Reduce to consider flips of Asn, Gln and His sidechains in addition to rotations of OH, SH and NH3 groups. One useful method of examining your hydrogenated structure is to use Mage and Prekin, located on the web site: http://kinemage.biochem.duke.edu
Reduce works with PDB format atom names. XPLOR is particularly cavalier about generating non-standard atom names such as _HB1 instead of the correct 2HB_ (the numbers are in different columns and the stereochemistry is backwards too). Thus you may need to use the pdbcns utility to convert them.
There are lots of parameters to Reduce, most of which you should not need, but if you are curious, reduce -help will list them.
See our description of all-atom contact analysis in:
Word, et. al., Visualizing and Quantifying Molecular Goodness-of-fit: Small-probe Contact Dots with Explicit Hydrogen Atoms, J. Mol. Biol. (1999) 285, 1711-1733, and
Word, et. al., Asparagine and Glutamine: Using Hydrogen atom Contacts in the Choice of Side-chain Amide Orientation, J. Mol. Biol. (1999) 285, 1735-1747.
e-mail:
J. Michael Word or
David C. Richardson
URL: http://kinemage.biochem.duke.edu
Biochemistry Department
Duke University
Durham, NC USA 27710