13c refinement and minimization

Dr. John Badger jb at MSI.COM
Wed Jan 7 15:12:08 EST 1998


The premature termination of the X-PLOR powell minimizer is
 a problem of which we are aware. This behaviour seems to occur
 most often when the energy term includes x-ray or NMR data.

 A workaround is to create a loop that breaks the minimization
 cycles into a series of small minimizations.

 For example
----
!
! Parameters for minimization
!
evaluate ($min_steps = 200 )
evaluate ($min_stage = 10 )



    evaluate ($min_count = 0)

    while ($min_stage > $min_count ) loop finalmin

      evaluate ($min_count = $min_count + 1)   

      minimize powell
        nstep = $min_steps
        nprint = 25
      end

    end loop finalmin
----

 would do 2000 steps of minimization as 10 x 200 steps
 of minimization.

 You could also use this strategy to put your own stopping
 criteria into the loop if 2000 steps seems like overkill.

 John Badger

 Product Development Manager, Macromolecular Crystallography
 Molecular Simulations



More information about the X-plor mailing list