Hi Bernard,
When I wrote my comment, I considered the same, since correlated
alternative conformations are in my experience quite common. However, I
do not believe it is the case here. The Phe and Tyr can have two
conformations: an "in" conformation, pointing towards the other, and an
"out" conformation, pointing away from the other. There is no problem
for both side chains occupying the "out" position, but only one of them
can occupy the "in" conformation, since both side chains in this
conformation would clash. This means that the sum of the partial
occupancies for the "in" conformation can not be greater than one,
which, as I recall correctly, is also what is observed with ~0.3 for the
"in" conformation and ~0.7 for the "out" conformation. However, the
partial occupancies need not be the same for the Phe and the Tyr.
My two cents,
Herman
PS: I was told that in refmac clashes between parital occupancies are
ignored if their sum is less or equal than 1.
________________________________
From: CCP4 bulletin board [mailto:[email protected]] On
Behalf Of Bernhard Rupp
Sent: Friday, July 27, 2012 5:02 PM
To: [email protected]
Subject: Re: [ccp4bb] Alternative conformation of Phe and Tyr?
Hmmm...I have a question for the experts: It seems to me from
the figure (although that is for Tri Ngo to verify), that
these partial occupancies are correlated. When conformation A of
TYR (in the green density) is occupied,
Phe B in the blue density probably cannot be occupied because of
a VdW clash. (Again, caveat that I can't see that exactly from the
static image)
I recall that Shelxl allows to constrain such occupancies, but
how are Refmac and Phenix dealing with that? How would I implement that?
Are clashes between partial occs just ignored?
Even if I am wrong about the existence of a clash in this
particular case, such situations are not uncommon imho.
Thx, BR
From: CCP4 bulletin board [mailto:[email protected]] On
Behalf Of Pavel Afonine
Sent: Friday, July 27, 2012 7:17 AM
To: [email protected]
Subject: Re: [ccp4bb] Alternative conformation of Phe and Tyr?
Hi,
you can't expect to see something with occupancy ~0.3 at the
same cutoff level as you use to see fully occupied sites. So most likely
the solution is to use a lower cutoff levels, as many suggested already.
Two maps you attached look totally expectable to me.
Pavel
On Fri, Jul 27, 2012 at 1:01 AM, Tri Ngo <[email protected]>
wrote:
Dear ccp4bb,
I am working on refinement of an enzyme structure. I notice an
alternative conformation (gauche+ and gauche-) on both residues Phe and
Tyr.
Here is the image of electron density before refinement. Based
on the different map intensity, I am quite sure there is an alternative
conformation on both positions.
http://i50.tinypic.com/29zaoap.jpg
However, after running occupancy refinement by Phenix, I don't
see the density map at the gauche- conformation of Tyr residue. Please
take a look at the image 2.
http://i50.tinypic.com/anzfrn.png
So do you think the gauche- conformation of Tyr exists in this
structure. Should I keep it in the final model because there is no
positive map there?
Thank you very much for your input.
--
Ngo Duc Tri
PhD Student
Structural Biology Lab
School of Medicine - Sungkyunkwan University - Korea
Phone: 031-299-6150 - Cell: 010-7774-8210
