 |
 |
Dear All, As suggested, I tried several things. I am getting MR solutions with TFZ > 8.0 (e.g. 12, 16) and high LLG (e.g. 150, 160). However, Rw/Rf remains above 0.50 after first round of refinement. In some cases, after some manual model building, Rw falls up to 0.46, however, Rf remains above 0.50. Any suggestions for further model building. The data is of ~1.9 Angst processed in P1211. With best regards, Shankar On Mon, Aug 1, 2016 at 1:48 PM, <[log in to unmask]> wrote: > Dear Schankar, > > > > you could also try automatic packages like Balbes or Morda. However, it > might not be a bad idea to try a more rational approach as well. With 27% > sequence identity, you template may or may not have the same fold. With 27% > sequence identity AND a similar biological function (e.g. similar reaction > catalysed, binding to a similar receptor etc.), your bets are much better. > (Except cases with large conformational changes like antibodies and > calmodulin). > > > > If molecular replacement fails, you should also look very carefully in the > space group assignment and in case of ambiguity try all possible space > groups for your MR searches. > > > > Best, > > Herman > > > > > > > > *Von:* CCP4 bulletin board [mailto:[log in to unmask]] *Im Auftrag von > *Sivakumar N > *Gesendet:* Montag, 1. August 2016 08:30 > *An:* [log in to unmask] > *Betreff:* Re: [ccp4bb] Molecular replacement with template having low > sequence identity > > > > Dear Shankar, > > How do you appreciate this approach of generating an Ensemble of > superposed homologous structures as a MR search probe? I found the FUGUE > server helpful in pulling out the distantly related structural homologues > for a given query sequence. Also, a manual or an automated method of > refining or sieving the superposed homologues to cut out the residues that > fall in the loop regions and possibly other residues at loci that introduce > any noisy-correspondences between the superposed structures can be useful, > in order to arrive at a better RMSD values between the equivalent Cα atoms > of this hybrid MR search probe. > > Regards > > Sivakumar,N. > > > > On Sat, Jul 30, 2016 at 6:20 PM, Shankar Prasad Kanaujia < > [log in to unmask]> wrote: > > Dear All, > > Is it possible to solve the structure of a protein having template with a > sequence identity of 27%. If yes, what is the best possible method. > > Any program which can automatically give some clue. > > With best regards, > > Shankar > > > -- Shankar Prasad Kanaujia, Ph.D. Assistant Professor Department of Biosciences and Bioengineering Indian Institute of Technology Guwahati Guwahati - 781039 Assam, India Tele: 0361 258 2228 Fax: 0361 258 2249 Email: [log in to unmask] Homepage: http://www.iitg.ernet.in/spkanaujia/ Google Scholar: https://scholar.google.com/citations?user=Zt4JSNYAAAAJ&hl=en