Hello Charles, Hope you had a good thanksgiving. :-)
So, right now the noise levels I am using are : [taken from the paper]. NH : 0.26 HZ Ca-Ha : 0.58 N-C' : 0.10 Ca-C : 0.10 What about HN-C' [ amide hydrogen and Carboxyl]? Since its of the same atom type as Ca-Ha : I am using 0.58. Is that OK? I did not find papers mentioning that value. Thanks Santhosh On Mon, Dec 2, 2013 at 3:33 PM, Charles Schwieters <char...@schwieters.org>wrote: > > Hello Santhosh-- > > Sorry for the slow response- Thanksgiving was busy. > > > > > This is more of a conceptual question. > > If we are generating synthetic RDC's [from a ensemble of conformations], > how do > > we add noise to the system? [noise - observation errors] > > > > For NOE's the acceptable way is 10-25% Guassian noise. > > > > For RDC's I can think of two ways: > > 1). Have a Guassian noise of of 0.26 HZ for NH and other bond vectors > have > > different errors. > > > > This was mentioned in the paper, Amplitudes of Protein Backbone Dynamics > and > > Correlated Motions in a Small R/β Protein: Correspondence of Dipolar > Coupling > > and Heteronuclear Relaxation Measurements. by Dr. Clore and Dr. > Schwieters. > > > > Since the RDC's have different magnitudes, depending upon the > orientation, this > > model can potentially add huge noises to some RDC's while not affecting > some > > others. > > > > 2). Use 10% Guassian noise of the magnitude of the RDC. > > I have not seen this method used in the papers I have read. > > > > What is the accepted method of adding noise to synthetic data > > > > Method 1 more closely corresponds to errors measured in experiments, > so that would be the correct choice. > > best regards-- > Charles >
_______________________________________________ Xplor-nih mailing list Xplor-nih@cake.cit.nih.gov http://cake.cit.nih.gov/mailman/listinfo/xplor-nih
