alternate conformations

[Axel T. Brunger]

> Dear X-PLOR users,
> I encountered a problem trying to refine alternate conformations of ligands in
> my protein. I'am refining a trimer with the segid's M,N,O. In order to set up
> the alternate conformations for the corresponding pigments they are labeled
> with segid's P,Q,R for the first conformation and S,T,U for the second
> conformation.
> Therfore I want to exclude all interactions of the type P-S, Q-T, or R-U.
> According to the manual the following should do the job:
>
>  constraints
>   inter (segid M or segid N or segid O) (segid M or segid N or segid O)
>   inter (not segid S) (segid P)
>   inter (not segid T) (segid Q)
>   inter (not segid U) (segid R)
>   inter (not segid P) (segid S)
>   inter (not segid Q) (segid T)
>   inter (not segid R) (segid U)
> end
>
> And well, it works! The is one slight disadvantage: The CPU time increases
> dramatically! Instead of a couple of days for my 3000K slowcooling job its now
> just below 2000K after 5 days!
>



Try to group your conformations in order to minimize the number of interaction
statements:

 constraints
   inter (segid M or segid N or segid O) (segid M or segid N or segid O)
   inter (not (segid S or segid T or segid U) (segid P or segid Q or resid R)
   inter (not (segid P or segid Q or segid R) (segid S or segid T or resid U)
 end

The CPU-time consumption is related to the number of "interaction" statements.

Axel Brunger