Re: [sdpd] Objects in Espoir

[Armel Le Bail <armel...@fluo.univ-lemans.fr>]

>When using Espoir, is the movement of unit cells used to describe 
>objects within the main structure unit cell entirely independent of the 
>edges of this main cell? In other words, do I have to make sure that 
>a unit cell containing a structural fragment is of minimum size to 
>ensure maximum freedom of movement?

When a rigid fragment or molecule or object is used as a model
in ESPOIR, two cells have to be considered :
   - the cell A in which you want the model to be located by Monte Carlo
     moves (rotations, translations),
   - the cell B used for describing the original model (or Cartesian coordinates).

You have to choose a model that may fit inside the cell A,
logically. But the cell B has no importance.

The model never moves in cell B. The coordinates in B are
transformed in coordinates in A and then the object is rotated or
translated in A. During these moves, there may be inter-object distance
checking. But my opinion is that no distance checking may
allow more freedom for a random path. The fact that some positions
may correspond to interpenetration of fragments and impossible
distances is a problem only if this is the final proposition... You have
to let the program find a minimum Reliability R factor which has to
correspond to a physically realistic final position.
There are few chances to obtain a quite good R together with a
completely false model. But if this occurs (use 50 reflections
per fragment), the model will have to be rejected, of course.

Also, the program may fail to find a satisfying solution, either
because of inadequacy of the hypotheses (wrong model or space
group) or because not enough independent tests were performed,
or etc. Testing the program and training yourself by using known 
structures is recommended.

Hope the answer is clear enough,

Best,

Armel