In article <337CBE59.41C6 at sneezy.fhis.net>, barry at SNEEZY.FHIS.NET (Barry Schweitzer) writes:
> Here is my next protein newbie question:
>
> From what I have read, distances involving methylene or methyl
> protons should have their upper bounds increased by 0.5 angstrom.
> I was taking a look at the g_protein_noe.tbl example file in
> the tutorial directory and while many distances looked reasonable:
>
> assign (resid 52 and name HA)(resid 3 and name HB2) 4.0 2.2 1.0
>
> other distance contraints looked like this:
>
> assign (resid 5 and name HD1#)(resid 34 and name HB#) 4.0 5.2 4.0
> assign (resid 5 and name HD1#)(resid 52 and name HN) 4.0 3.7 2.5
>
> indicating that increased bounds were used for both lower and
> upper bounds, and that the magnitude of this increase was a function
> of the number of ambiguous protons. Is this common practice?
> What rules of thumb for calculating these bounds do others use?
A correct way of dealing with this is to follow the advice in:
Fletcher et al., (1996) J. Biomol. NMR vol 8 pp 292-310
In essence, they point out that if you use X-PLOR's "sum averaging" (which
is not averaging at all!), then you don't need to do *any* distance corrections
at all - the calculation of the "effective" distance for every restraint,
whether involving wild-card atom names or not, takes into account the
multiplicity of the contributions to the intensity. The benefits are obvious:
it's easier to do and you can be confident that you are using your restraints
as "tightly" as is justifiable.
Of course, any additional violations that you might end up with are your own
problem! The whole issue of calibrating your restraints is a can of worms
that I don't want to go into.
Andrew
--------------------------------------------------------------------------------
Dr. Andrew Raine, Department of Biochemistry, University of Cambridge,
Tennis Court Road, Cambridge, CB2 1QW, United Kingdom.
Telephone: +44 1223 333744, FAX: +44 1223 333345
email: A.R.C.Raine at bioc.cam.ac.uk
