[Chimera-users] updating hydrogen bonds in an MD movie

Eric Pettersen pett at cgl.ucsf.edu
Wed Oct 3 12:46:08 PDT 2007

Hi Eric,
	I have some things to supplement Elaine's answer.  There is a  
feature in Chimera's Python layer for speeding up H-bond analysis of  
trajectories that is currently only available to scripts.  The  
feature is caching the identification of the donors/acceptors of a  
structure so that it doesn't have to be redone for each frame.  I've  
now added code so that it is available in the hbond command as the  
"cacheDA" keyword (value should be true or false).  Empirically,  
turning caching on speeds up the trajectory playback by more than a  
factor of two.
	One way to get the feature is by downloading a test build newer than  
October 1 from http://www.cgl.ucsf.edu/chimera/alpha- 
downloads.html .  However, I'm aware of some other problems in those  
builds so I don't necessarily recommend that right now (the problems  
_should_ be fixed in a week or so though).  What you can do instead  
is edit Chimera's Python code to add the keyword yourself.  You would  
need to edit the file:

<your Chimera installation>/share/FindHBond/base.py

(on a Mac you need to control-click on Chimera.app and choose "Show  
Package Contents", then in the new Finder window maneuver to Contents/ 

You need to add the keyword to the createHBonds function and  
subsequent calls to findHBonds.  If you are using the 1.2422 release,  
change line 852 from:

	reveal=False, namingStyle=None, log=False,
	reveal=False, namingStyle=None, log=False, cacheDA=False,

line 902 from:

		interSubModels=interSubmodels, cacheDA=cacheDA)

and line 930 from:

			interSubModels=interSubmodels, cacheDA=cacheDA)

	With these changes, you can use the command "hb cache true" in the  
MD Movie scripting interface to show H-bonds using caching.

> Also, may I take this occasion to point out that requiring consecutive
> frames to have consecutive  numbers in the file names is inconvenient.
> For instance,
> my simulation is 41 ns long, and the structures are stored (in our
> in-house format) every 1 ps.  So rather then write out 41000 .pdb  
> files,
> I'll write them out every 10 or 100 ps.  To use chimera to view a  
> movie, I
> have to rename the files using a perl script.

The non-consecutive thing hadn't even occurred to me, so thanks for  
pointing it out.  I'm responsible for a lot of parts of Chimera so  
don't nearly as much time to work on MD Movie as I would like, so it  
may be some time before I am able to do anything to remedy this.   
I'll open a change-request in our bugs database with you cc'ed so  
you'll know when something happens


                         Eric Pettersen
                         UCSF Computer Graphics Lab
                         pett at cgl.ucsf.edu

More information about the Chimera-users mailing list