Zheng Zhou wrote:
Hi, Ed
I am dealing the similar problem. I checked CNS qindividual.inp. But how
do I refine one compound with two or more possible conformations (mainly
due to one bond rotation), each of wihich has a different occupancy?
Thanks in advance.
Hi Zheng,
Others can answer better than I, but for what it's worth:
You have basically two methods for refining alternate,
possibly overlapping, models.
One is the "alternate conformations" formalism which I have
not used but is documented in the CNS FAQ
http://cns.csb.yale.edu/v1.2/faq/text.html#xtal_refine
(search for:
Q. How do I deal with alternate conformations in my refinement? )
The other way is by modeling both structures (with different chain
letter or range of residues or something) and turn off vdw
interactions between the two entities which never occur
simultaneously in the same unit cell using the igroup statement.
Some hints can be gleaned from the answer to this different question
(in the same FAQ):
Q. My structure contains a molecule which lies across a symmetry operation. This means that some parts of the molecule are mapped onto each other by symmetry. It is not a
special position case, because no one atom lies on the symmetry operation exactly. How can I tell CNS to refine the molecule while allowing the overlap to occur?
A. You can use the igroup statement to turn the pvdw (Packing-Van-der-Waals)
interactions off by setting the weight to zero:
-------------------
Syntax for the statement is something like: interaction (selection1 selection2)
means for it to check interaction between each atom in selection1 with
each atom in selection2. What I am using (with CNS 1.1) is:
igroup
interaction ( &atom_select and not (segid "M" or segid "Z" or attr store9 >
0))
( &atom_select and not (segid "M" or segid "Z" or attr store9 >
0))
interaction ( &atom_select and not (segid "E" or attr store9 > 0))
( segid "M")
interaction ( &atom_select and not (segid "R" or attr store9 > 0))
( segid "Z")
{
interaction ( segid "E ") ( segid "M ") weights * 1.0 pvdw 0.0 end
interaction ( segid "R ") ( segid "Z ") weights * 1.0 pvdw 0.0 end
}
evaluate ($alt=1)
while ( $alt <= $nalt ) loop alcs
interaction ( &atom_select and ( attr store9 = $alt or attr store9 = 0 ))
( &atom_select and ( attr store9 = $alt ))
evaluate ($alt=$alt+1)
end loop alcs
end
Here segid M and Z are the same subunits as E and R, but in alternate
conformations.
Store9 has to do with the formally declared alternate conformations (see "atom
selection").
The first interaction statement ignores M and Z and checks interaction of
everything
else with everything else. The second checks interaction of everything in M with
everything except its alternate self E; the third checks Z with everything
except R.
I see I have {commented out} the section that sets pvdw weights to zero between
alternate conformations of the same thing, but I'm not sure if this is right.
It may be this is not needed since we avoid checking those interactions, but I
think
at one point we were turning off the NBONDS messages but not the vdw
interaction,
so there may be two separate things needed here. I would be glad for any
clarification.
Ed
On Dec 17, 2007 2:24 PM, Edward Berry <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
I think the correlation between occupancy and B-factor depends
also on the size of the ligand (relative to resolution).
Bob Stroud, I think, has estimated occupancy by comparing
the integrated electron density of the ligand with that of
a well-defined, isolated water (assumed to be at unit occuancy?).
In principle the integrated electron density is not affected
by applying a B-factor, it is just spread out over a wider
area. In the case of a single atom at 3 A resolution, it
is spread out under the neighboring atoms and effectively
lost, so it is hard to distinguish high B-factor from low
occupancy.
In a large ligand most of the atoms are inside the ligand,
so their spread-out density remains inside the ligand
and gets counted in the integrated density. In that case
high B-factor has a very different effect than low occupancy,
as only the latter reduces the total electron density of
the ligand.
During a previous reincarnation of this thread I did the
simple test of refining occupancy and B-factor for a
stretch of the protein (holding the rest of the protein
at unit occupancy) in CNS 1.1, and I felt the results
were quite satisfactory (don't have the specifics now).
Ed
Anastassis Perrakis wrote:
>> I have already changed occupancies as Eleanor mentioned, and got
>> approximate values. But my hope is to try to get much precise
ones if
>> possible.
>>
> I never expected to preach the 'Kleywegt Gospel' in the ccp4bb,
> but in this case what you need is more accurate answers, not more
> precise ones
> (or better both, but precision alone can be a problem, and you can
> easily get
> 'precise' but inaccurate data easily by making the wrong assumptions
> in your experiment)
>
> http://en.wikipedia.org/wiki/Accuracy
<http://en.wikipedia.org/wiki/Accuracy>
>
>> I have heard from my colleague SHELX can refine occupancies, and
>> got its license. I'll next try SHELX.
>
> I think that phenix.refine can also do occupancies ?
> The problem is not if the program can do it, but if at your
specific case
> you have enough information to do that in a meaningful way.
>
> For a soaking experiment and 1.5 A data, I would say that Eleanor's
> suggestion
> of tuning Occ based on B, is as close as you would get, accurate
enough
> given the data,
> although not necessarily too precise.
>
> Tassos
