*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
On Thursday 05 October 2006 06:49 am, Nalam, Madhavi wrote: > > I am using TLS refinement followed by restrained refinement. Is there > a minimum number of the restrained refinement cycles to be run after > the TLS refinement? In other words, I run 5 cycles of TLS refinement > and then 2 cycles of restrained refinement (From the 3rd cycle, the > Rfree starts going up). Is it correct if I stop the refinement after > two cycles of restrained refinement? You have failed to tell us whether your "restrained refinement" is being done on coordinates only, or coordinates plus Biso values. The default protocol when you run refmac via the ccp4i interface is to follow up on the TLS refinement by restrained refinement of both coordinates and individual Biso values. However, at low resolution you may well do better to use a "pure" TLS model: - start with uniform Bs - refine TLS - subsequent refinement of coordinates only (no Biso) This will reduce the number of parameters being refined, and hence reduce the chance that you are over-fitting your data. At low resolution (let's say worse than 2.5A) this can be a major concern. -- Ethan A Merritt Biomolecular Structure Center University of Washington, Seattle WA
