On Thu, Oct 22, 2009 at 11:41 AM, vanessa delfosse < [email protected]> wrote:
> I'm looking for people using Phenix to refine RNA or DNA structures. I work > on RNA structures since some months and I observed quite large differences > between CNS and Phenix, especially concerning rms deviations. > A quick search in the PDB suggests that people working on nucleic acids > prefer to use CNS. Does it really better for nucleic acid refinement than > Phenix ? > You will probably have better luck asking this on the Phenix mailing list: http://www.phenix-online.org/mailman/listinfo/phenixbb I can't speak for nucleic acids crystallographers, but phenix.refine has only been around for a few years, so surveying the PDB isn't going to be very informative. CNS used to be the only refinement program that did torsion angle dynamics, which is particularly useful for nucleic acid structures at low resolution. The latest version of Phenix (1.5-2) will do this too, however. There are many more less obvious differences between the two programs that would explain the large differences in results - although, if by "rms deviations" you mean deviations from ideal geometry, this is usually means you need to adjust the X-ray/stereochemistry weighting.
