Crystallography‎ > ‎Phasing‎ > ‎

Molecular Replacement

FROM phaser/RMS and sequ ID:
RMS = max(0.8,0.4*exp(1.87*(1.0-ID)))  {ID is the fraction identity}. 
"The RMS deviation estimated from ID may be an underestimate of the true value if there is a slight conformational change between the model and target structures
Sequence IDRMS deviation
100%0.80Å
64%0.80Å
63%0.799Å
50%1.02Å
40%1.23Å
30%1.48Å
20%1.78Å
--> limit 0%2.60Å

From
ccp4bb, general advice:
Run mindless refinement straight after MR. Check that R-work and R-free both decrease - if not there is a good chance it isn't a solution. Sort the resultant coordinates on B value (sort -n +10 -11 output.pdb). Check which bits have the highest B factors. It usually finds loops. Set all the occupancies of those loops to 0.0, also set some good side chains ocs to 0.0. Refine again, look at maps - have the good bits re-appeared?

Also from ccp4bb, how to prepare models for MR:
Run the sequence against all PDB using:
2. http://www.ebi.ac.uk/Tools/fasta33/index.html (using "Protein structure Sequence" from the box "Databases")
Both 1 and 2 run the same search but give a different output.
3. Also try http://meta.bioinfo.pl/submit_wizard.pl (it will predict 3D structure using many methods)

Links:

From?

"if you were working on the ribosome, then MR at 8.5A would probably be straightforward.  But what is important is the ratio of the size of the molecule to the resolution limit, and for molecules of a typical size this resolution is almost certainly too low.  It basically comes down to questions of how much information you have, which you can think about in terms of how distinctive the shape is at your resolution or even how many reflections you have"