[Ifeffit] Atoms restraints

[Matt Newville]

Dear Max,

> I'm a new user of the Feffit package and I have a question about
> atom restraints for the fitting.
> 
> I'm dealing with metal binding protein, and somehow I need to
> keep some atoms in a fixed structure (mainly histidine rings)
> while leaving a certain degree of freedom to the remaining atoms
> of the environment.
> 
> I'd like to know if there is a library containing histidine
> structure restraints or if exists any input file of a similar
> matter.
 
I don't know of any such library, though such constraints and
restraints could be done -- at least in principle. (I)Feffit is
very strongly tied to the Sum-Of-Paths approach to EXAFS, which is
not particularly good at dealing with groups of atoms as aingle
entities.

I believe this would be a significant amount of work.  Since
(I)Feffit uses a sum-of-paths and since paths (in the form of
Feff's feffnnnn.dat files) do not necessarily come from a single
cluster of atoms (this is by design and is a _benefit_ in many
cases, though not here!), it will not necessarily be easy to
identify the x,y,z coordinates of the individual atoms in a ring,
let alone know that certain atoms belong together.  The path to
each atom would have to be identified by hand.  Then, the
mathematical constraints keeping the distances and angles between
atoms within the ring fixed while allowing the group to move
relative to the central atom would have to be worked out in terms
of the allowed motions.  To be honest, I haven't really thought
about how to do this in any detail.

--Matt