IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

Slow convergence

Brent Segelke segelke1 at LLNL.GOV
Fri Sep 4 10:40:20 EST 1998


At 12:49 AM 9/4/98 -0700, you wrote:
>On 3  1998, Jan Dohnalek wrote:
>
>> Hi All,
>> 
>> Has anyone had this experience of VERY long convergence
>> of ~1.8A positional refinement? Usually 1000 cycles or more.
>> Is it really WRONG? Or is it just long ....?
>> 
>> Rfactors about 20.0 and 27.0.


> 
>why not use slowcool SA protocol?
>It converges more quickly.


I would recommend against this.  In my experience, if you use SA for
moderate resolution or low resolution structure determination, you have an
ill behaved FreeR; and if you have good, high resolution, data, SA is
unnecessary (you may still get worse FreeR's).  'course there are lots of
caveats but as somebody famous (Oded Livnah) once taught me, you need a
license to use SA.
-----------------------------------------------------------------
-----------------------------------------------------------------
If it crystallizes, it's important!
-----------------------------------------------------------------
-----------------------------------------------------------------
Brent Segelke, Ph.D.                         Phone (510) 424-4752
Lawrence Livermore National Laboratory       Phax  (510) 422-2282
Macromolecular Crystallography               
LLNL-BBRP L452				    
POB 808                                      
Livermore, CA 94551                       email segelke1 at llnl.gov
URL             http://www-structure.llnl.gov/cvs/Brent/brent.htm
-----------------------------------------------------------------



More information about the X-plor mailing list

Send comments to us at biosci-help [At] net.bio.net