 |
 |
Shankar Prasad Kanaujia wrote: > Dear all, > I am refining a twinned (twinning fraction = 0.213 from Yeates' method) > data using CNS. I refined in two ways. > 1. Firstly, I created the .cv file using make_cv_twin.inp from CNS and > then e.g. used minimize_twin.inp for minimization. The Rwork and Rfree > were 19.55% and 22.06% respectively. > 2. In the second case, I detwinned the .cv file (.cv file created in the > first case) using detwin_partial.inp from CNS. And then e.g. used > minimize.inp for minimization (Here, I am refining as there is no > twinning). The Rwork and Rfree were 24.5% and 26.59% respectively. > > In both cases, I used the same model and without any water. The > resolution of the data is 2.0 Angs, space group R3, twinning operator > (h,-h-k,-l) > > By seeing the Rwork and Rfree, I can guess that first case is better. I > also superimposed two minimized pdb's and RMSD = 0.123 Angs. > > Now My question is which of the above two ways is the right way to refine > the model. And why there is almost 5% difference in R-factor values. > > Any suggestions or references will be highly appreciated. > Thanking you all in advance. > > -- > Yours Sincerely, > Shankar Prasad Kanaujia > IISc, Bangalore - 12 > Mobile: 9845631581 > > > Dear Shankar, the detwinning procedure increases the errors of the observed F's. This is particularly dramatic for twinning fractions close to 0.5. This probably explains the higher Rvalues using detwinned data. Twinned data have a tendancy to look quite good concerning the Rmerge, as the proportion of weak reflections is smaller. For similar reasons the refinement against twinned data appear to yield relatively low R-factors. I think in the end both of your refinements will give about the same result and the real R-factor would be in the middle of the two values obtained by the different techniques. Greetings Wim Burmeister -- ******************************************************************************* Wim Burmeister Professeur, Membre de l'Institut Universitaire de France Unit of Virus Host Cell Interactions (UVHCI) UMR5233 UJF-EMBL-CNRS 6 rue Jules Horowitz B.P. 181, F-38042 Grenoble Cedex 9 FRANCE E-mail: [log in to unmask] Tel: +33 (0) 476 20 72 82 Fax: +33 (0) 476 20 94 00 http://www2.ujf-grenoble.fr/pharmacie/laboratoires/gdrviro *******************************************************************************