Dear Charles, Thanks for your advice. My starting structures dose not give a good fit to RDC (PALES SVD simulation only gave regression of 0.75). I do not have a protocol of refining Da, Rh using gridsearch as described by Clores et al in J. Magn. Reson. 133, 216-221 (1998). Could you please advice me where to get this protocol? I did comment-out the lines referring to rdcs when using anneal.py without RDC, somehow the calculation is very slow compared with that using RDC. Maybe I made a mistake somewhere. I will attach the anneal.py for you to check later since I am out of town and cannot login to my Linux account. I appreciate your hepful suggestion!
Happy Easter Holiday! Best wishes, Hongyan Quoting Charles at Schwieters.org: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > Hello Hongyan-- > > > The problem for me is to choose which program to refine my protein > > structures using RDC. Xplor-NIH anneal.py script did gave rise to > > well-converged structures but there seems no much differences with or > > without RDC, both in RMSD and structures. However, in Aria, I do see > > the differences when RDC is used (lower RMSDs and slightly different > > helical angles). When I used anneal.py to calculate structures without > > RDC, I just input a empty rdc.tbl, not sure if > > it is OK. > > that should be ok. Better to just comment-out the lines referring to rdcs > > > Since when I added # to those related to tensor, rdc etc, running > > anneal.py tooks much longer time. I don't konw why. > > not exactly sure what you mean. It does take longer when rdcs are > included. > > > > > My people did use the method as described by Clore et al to refine Da and > Rh > > even if they have an initial structres (since they don't fit well with RDC > > data). I think the old-verion of Xplor-NIH could not float tensor, how > > to get a good set of Da and Rh is also important issue. > > anneal.py requires a decent approximation to Da and rhombicity- this can > be gained by the approach in J. Magn. Reson. 133, 216-221 (1998). > > If you have good starting structures you can use protG/calcTensor.py (or > protG/grid_search.inp if you want to wait longer) to determine values of > Da and Rh. Otherwise, please use the above reference. Once you have a > good initial guess at Da, Rh, anneal.py and refine.py can fine tune > things. > > hope this helps-- > Charles > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.6 (GNU/Linux) > Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> > > iD8DBQFGE5IkPK2zrJwS/lYRAk/rAJ9ZS9NpD5bh58caM9b6uCbTmX6g9gCfe7Sy > 7RwxFZQ3CrvqGGobnxZ9GSU= > =OU0R > -----END PGP SIGNATURE----- > Dr. Hongyan Li Department of Chemistry The University of Hong Kong Pokfulam Road Hong Kong
