You should not worry. A Ramchandran outlier does not mean a residue must never
be in such a configuration, but that it is energetically unfavourable. Since you
describe that the chemical environment supports the respective conformation it
tells you that these residues have a different net energy (different from what
one might expect from an "undisturbed" residue).
This applies for your 12-14 outliers, too: when their conformation energetically
makes sense, do not worry.
(Somebody firmer in statistics might be able to tell you how many "outliers" you
SHOULD expect, and I suppose with as many as 1200 residues you are not only
chemically safe (for the above reasons), but also statistically on "the safe
side".
I would say that 2.1A is good enough to believe in your observations.
Tim
On Fri, Apr 30, 2010 at 08:27:27AM -0700, xaravich ivan wrote:
> Hi everyone,
> I have a few residues sharing hydrogen bonding and salt bridges and they
> tend to be fitting the density perfectly, but are Ramchandran outliers,
> irrespective of the cycles of refinement.
> It seems to me that it might be expected as these residues have some kind of
> strained geometry because of the H-bond/ salt bridges.
>
> Am I correct or should I worry.
>
> Also I have around 1200 residues in my structure and there are about 12-14
> outliers (including those mentioned above) which would not buzz from their
> spot.
> Should I really worry to get the number of ouliers much below 1% to get the
> reviewers nod. What is a good acceptable percentage of outliers in a
> structure that is otherwise reasonable?
> The Rfactor/Rfree is 23/29 at 2.1 Angs resolution.
>
> Thanks,
> Ivan
--
--
Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen
GPG Key ID = A46BEE1A
|