Tom,
Welcome to the protein crystallography community. I hope you take all
the teasing in good humor. And a few protocol lessons about sharing
data, etc. By now you have discovered that structure factor files and a
possible MR model are like catnip to structural biologists. I learned
this when I solved my first structure while on a sabbatical leave. I ask
one question about a tricky (to me) MR solution, share the structure
factors with ONE postdoc, and the next thing I know the whole building
is solving my structure. (Cover ears and go La La La La La.) You have
taken this to a whole new level! :)
To echo the advice of many others, nothing beats actually looking at the
solution with a critical eye:
1. is there sensible electron density around the model? (You may want
to modify the model to poly-Ala or truncate to the nearest similar
residue--you can do this using CCP4 programs.) If sequence identity
is really poor a poly-Ala search model may be better. Once you get a
preliminary solution, I've used Parrot and Buccaneer successfully to
autobuild most of the structure from a relatively poor starting
model. But there is nothing wrong with manual rebuilding from a good
starting point with high homology.
2. Does the solution pack well, with definable solvent channels and
reasonable protein-protein contacts? (Use symexp in Pymol or turn on
symmetry molecules in Coot) Sometimes it is easy to discover a
missing or "extra" molecule in the ASU this way. That's what
happened with my first structure solution--I thought I had 4 chains
in the ASU and it was really 6, which was quite obvious from the
packing of my partial (as it turned out) MR solution. An extra two
chains fit perfectly in the "hole" in the packing view.
3. Don't forget to do a quick Matthews coefficient calculation before
MR to get some idea of how many molecules are in the ASU. It's not
perfect, especially for high copy numbers, but it's a start. It's
normally easy to tell 1 from 2 molecules in the ASU. Not so much 6
vs. 8.
4. It's not unusual for raw MR solutions to start at R>40%. But if you
are on the right track, the initial refinement in Refmac should
drive that down to 30% or lower pretty quickly. If R doesn't drop
rapidly, you likely have an issue with the MR solution. Be aware
that in the latest versions of Refmac, that first refinement may
take as many as 20 cycles to converge, not the default 10 cycles.
After that, 20 cycles is usually overkill.
Have fun, and solve your protein structure.
Cheers,
_______________________________________
Roger S. Rowlett
Gordon & Dorothy Kline Professor
Department of Chemistry
Colgate University
13 Oak Drive
Hamilton, NY 13346
tel: (315)-228-7245
ofc: (315)-228-7395
fax: (315)-228-7935
email: rrowl...@colgate.edu
On 3/27/2013 12:22 PM, Tom Van den Bergh wrote:
Dear members of ccp4bb,
I need some help with the refinement of my structure of a variant of
mRFP (monomer red fluorescent protein, sequence in attachment). I have
done molecular replacement with phaser with model 2VAD of protein
database. Then i have done some model building phenix.autobuild. (2
pdb's (overall...), freeR flags and log file attached) When i refine
with phenix.refine my structure i get a R-value of 0,42 which is still
way too high. (redfluorescent protein.pdb, .mtz and logfile attached)
When i look at the structure in coot i find many unmodelled blobs and
many outliers in density analysis and rotamer analysis. The problem is
that there are so many problems with my structure, that i dont know
where to begin. Could you try some refinement for me, because this is
first structure that i need to solve as a student and i dont have too
many experience with it.
Greetings,
Tom