Atomic model building by docking.

roseman at mrc-lmb.cam.ac.uk roseman at mrc-lmb.cam.ac.uk
Fri Dec 6 02:06:13 PST 2002



Dear 3D EM list members and colleagues,

I am pleased to announce a new - much faster - release of DockEM, my
program for building atomic models of macromolecular complexes by
fitting/docking the structures of domains into EM density maps.

Extensive searches of complete volumes can now be done in under 24 hours on
a (single) standard workstation. These are complete exhaustive 6 parameter
searches of the rigid body models of domains in the entire volume of the
map of the macromolecular complex. 

For example; for a 12 A map of GroEL (or the GroEL-GroES complex) 
sampled at 4 A/pixel the full search of the 64 cube map will take ~18 hours 
on a DEC alpha workstation with a 600 MHz EV6 processor. That is sampling the
angular space in 4 degree intervals. A finer angular search of 2 degrees
that is restricted to a 32 cubed region (about 3 asymmetric units) will
take ~15 hours. 

The program is still essentially as described in the original publication
(Roseman 2000, Acta Cryst D56, 1332-1340.), but the new version takes
advantage of a Fourier algorithm for implementing the real-space
correlation, that I have recently developed.  For details see my
Ultramicroscopy manuscript, in press, available online 29 October 2002.
This manuscript describes a 2D application of the fast local real-space
correlation function for particle picking from electron micrographs. See
articles in press, Ultramicroscopy, science direct,
(http://www.sciencedirect.com)

Please contact me if you would like an update or you would like to use the
program.

Yours sincerely,

Alan Roseman.


-----------------------------------
Research Associate
MRC Laboratory of Molecular Biology
Hills Road
Cambridge
England
CB2 2QH
roseman at mrc-lmb.cam.ac.uk



----------------------------------------------------
To unsubscribe from this e-mail list, send mail to
majordomo at sdsc.edu
with the words
unsubscribe 3dem
in the body of the message.
Do not send unsubscribe messages to the entire list
-----------------------------------------------------



More information about the 3dem mailing list