On 17 March 2013 13:19, Robbie Joosten <[email protected]> wrote:
Small addition to Ian's comment. The value you give with 'weight auto
$value' is a starting value. Refmac will gradually change it if needed
(it's autoweighting after all) and your starting value does matter
somewhat. Based on Ian's advice PDB_REDO uses a starting value of 2.50
which seems to do the trick most of the times.
Cheers,
Robbie
Good point, here's an example from a Refmac log file of that happening in
practice using Refmac_5.8.0031 (output redacted so not showing every
iteration):
Data line--- WEIGHT AUTO 0.95
Weight matrix 1.00564407E-02
Actual weight 0.94999999 is applied to the X-ray term
Weight matrix 1.19658876E-02
Actual weight 0.94999999 is applied to the X-ray term
Weight matrix 1.67442132E-02
Actual weight 1.2349999 is applied to the X-ray term
Weight matrix 2.25313306E-02
Actual weight 1.6054999 is applied to the X-ray term
Weight matrix 2.31216233E-02
Actual weight 1.6054999 is applied to the X-ray term
Weight matrix 2.32296251E-02
Actual weight 1.6054999 is applied to the X-ray term
Weight matrix 2.34125387E-02
Actual weight 1.6054999 is applied to the X-ray term
Weight matrix 3.04648653E-02
Actual weight 2.0871499 is applied to the X-ray term
Weight matrix 3.05170882E-02
Actual weight 2.0871499 is applied to the X-ray term
So here I inputted Wa = 0.95 (I deliberately chose it too small), and
Refmac calculates what the equivalent value of Wm (1.00564407E-02) would
have been (i.e. inputting WEIGHT MATRIX 1.00564407E-02 would have given
identical results to WEIGHT AUTO 0.95). However what is actually used
('Actual weight') in the refinement calculation is Wa. Then Wa is
automatically optimised according to Refmac's internal criteria, giving a
final optimised value Wa = 2.09. Note that Wm changes as the model is
improved even when Wa is kept fixed.
Cheers
-- Ian
>
