Weird blob appears

From: Jacob Keller
Date: 27 October 2011 17:27


Dear Crystallographers,

In the course of a reasonably smooth refinement, all of a sudden there
is a huge worm-hole-type blob in the electron density (see pics). Has
anyone seen this before? Is it some effect of the refinement
over-fitting or something? For some background, resolution is 2.2Ang
and there are 4mol/ASU, but one of them is markedly worse in terms of
density than the rest, and appears to be in a non-crucial part of the
lattice, as if just sitting in one of the interstices. R/Rfree is
~19.5/24.5.

Jacob



----------
From: Rafael Molina


Maybe sugars?

El 10/27/11 6:27 PM, Jacob Keller escribió: -----------------------------------------------------------
Rafael Molina, PhD

Structural Biology and Biocomputing Programme

CNIO - Centro Nacional de Investigaciones Oncológicas
C/ Melchor Fernández Almagro, 3, E-28029 Madrid (España)

www.cnio.es<http://www.cnio.es>
-----------------------------------------------------------



----------
From: David Briggs


I agree with Rafael,

From those pictures it looks like a sugar chain - maybe 2-3
saccharides in a row.

HTH

D
============================
David C. Briggs PhD
Father, Structural Biologist and Sceptic
============================

----------
From: Jacob Keller

Part of what is bothering me is that the density showed up at one
exact point in the refinement, and I am currently testing what exactly
it was that changed things. To me, the blob almost looks like a mask
of the molecule, and there is very little 2Fo density, so that's
weird. I am also really bothered that the R value is so low, and yet
there's this huge blob. Something seems buggy to me. Also a
clarification: the blob runs into a two-fold axis in the middle, so
things get dodgy there too. I will let you know what I find out.

JPK

----------
From: Yuri Pompeu


Jacob,
By simply looking at the figures you show, it does look like you have some type of long, maybe polymeric, molecule bound.
With that being said:
1- It is in the symmetry axis so maybe be a little noisy there
2-If you are in doubt about it being real or not check the density and how it fits into your protein and (symm. related neighbours of course). If the desity appears to be in position for some real interactions- i.e, there are some peaks at H-bond distances of polar groups etc...- it may be real. If its all random and through your protein atoms, probably not
Keep in mind alternate conformers....
HTH

----------
From: Jacob Keller


Blob is gone--something funny happened, I guess. I went back to using
the original mtz from scala, removed and replaced a bunch of waters,
and no more worm! I can't really figure it out, and wish I knew
exactly what happened, but I think I am just going to non-chalantly
move along.

Jacob

----------
From: Artem Evdokimov

Weird. It's pretty much exactly what a nonbranched sugar (triose, in this case) would look like.

 

were the dimensions consistent with three sugars? were there any collisions with backbone or side chains inside/near the mystery density?

It really looks like maltotriose... is the protein a sugar binder by any chance?

 

Artem

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

Hello Jacob,
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^ maybe that solves your question:
You should always use the same file as input, i.e., should never have
used anything but the original mtz from scala...?
Tim

> [...]
- --
- --
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen



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


Curious.  What were you using - the refmac output mtz?  Just for the
record - the refmac output mtz contains *modified* amplitudes, and Garib
said many times it should not be used as the input for the next cycle...
However, from what I understand about scaling applied to output fobs, it
is highly unlikely that it will generate an extra density of the kind
you observed.


--
Oh, suddenly throwing a giraffe into a volcano to make water is crazy?
                                               Julian, King of Lemurs