Saturday, 21 January 2012

model building at 3.2 A

From: atul kumar
Date: 8 December 2011 06:47


Dear all

I have crytals which diffract up to 3.2 A at synchrotron, I am solving this by molecular replacement. I have built 70% of the model successfully,but the problem is it have a very poor density for some 30-40 residues at N terminal. I can't build anything in this region,this could be because of disordered structure or because of low resolution.what are the things which I can try to improve this?

suggestion are requested!

thanks
regards'
Atul Kumar


----------
From: Tim Gruene

Dear Atul,

the "resolution" of your data set would affect the whole unit cell and
not just the region of 30-40 residues. The more likely reason is a
disordered N-terminus. There is nothing to do about this in silico.

It may be worth recloning your protein with that N-terminal part chopped
off - there is a fair chance that the product is more stable and hence
diffracts much better than 3.2A.

Regards,
Tim
- --
- --
Dr Tim Gruene
----------
From: Francis E Reyes


On the other hand, shooting a lower resolution crystal may get you the conformation of the disordered domain.

Surprising at first thought, but was true in p97/VCP from the Brunger Lab.


F

----------
From: Edward A. Berry


Hmm- I wonder if B-factor unsharpening (applying a large POSITIVE B-factor to the data)
would have the same effect?

----------
From: Ed Pozharski


Both, and also the presence of disordered fragment may be the reason the
resolution is low, thus the already mentioned suggestion to remove it
(unless, of course, that renders your protein inactive and thus
compromises your structure-based conclusions).
If the region in question is disordered in solution, pretty much
nothing.  Look for a different crystal form.

Having a disordered region does not make your structure invalid.  In
fact, it may be important for understanding the biological role of the
protein.


--
"I'd jump in myself, if I weren't so good at whistling."
                              Julian, King of Lemurs

----------
From: Savvas Savvides

Dear Kumar
Could you please provide some information on your refinement protocol/progress and quality of your partial model thus far. Also, what is the quality and completeness of your data, in particular in the lowest resolution to 10 angs range.   In this way one might be able to provide more concrete feedback.
Experience from several structures around such resolution tells me that it is very much worth spending time optimizing what you can reliably fit and refine. This approach will pay off handsomely in terms of map quality and might lead to modeling what now may seem uninterpretable.
The latest Phenix and Refmac versions are great, and their B-sharpening protocols can reveal a lot. In addition Buster with its handling of missing domains in Fc can also help a lot.
Did you try any phase improvement by density modification? No. of molecules in asu? Also getting some experimental phase info (even at low resolution) might also help.

Best regards
Savvas


No comments:

Post a Comment