Hi Vandana,
You already have some excellent suggestions from the experts. I agree with Randy that high sequence identity does not mean a good MR search model and sometimes models with lower sequence identity may give MR solutions. We recently faced a similar problem where classical MR (using search models with significant sequence identity) was unable to give any solutions, but eventually the structure was solved using phenix.mr_rosetta (http://journals.iucr.org/f/issues/2016/06/00/hv5326/index.html).
Good luck!
Avinash
|