[phenixbb] refine anomalous

Andreas Förster docandreas at gmail.com
Mon Sep 8 00:22:33 PDT 2014

Dear all,

I've solved a structure from a SAD dataset collected at 1.54 Å.  The 
structure contains a good number of sulfurs and also what I think is a 
potassium ion.

To verify, I chose a refinement strategy including "Anomalous groups". 
This worked well in an important way - R values down by 2 percentage 
points - but the refined f' and f" values are way off.

I start out with these parameters in *.eff:

     anomalous_scatterers {
       group {
         selection = (name SD or name SG)
         f_prime = 0.319342
         f_double_prime = 0.557592
         refine = *f_prime *f_double_prime
       group {
         selection = name K
         f_prime = 0.365562
         f_double_prime = 1.06736
         refine = *f_prime *f_double_prime

What I get back is:

Anomalous scatterer group:
   Selection: "(name SD or name SG)"
   Number of selected scatterers: 15
   f_prime:        -1.15797
   f_double_prime: 0.00577091
   refine: f_prime f_double_prime
Anomalous scatterer group:
   Selection: "name K"
   Number of selected scatterers: 1
   f_prime:        -4.41689
   f_double_prime: 0.00271946
   refine: f_prime f_double_prime

I thought I'd get close to the book values for S and would be able to 
identify the other anomalous scatterer by the refined f'/f" values. 
Naiveté or stupidity?

Thank you.


                   Andreas Förster
      Crystallization and X-ray Facility Manager
            Centre for Structural Biology
               Imperial College London

More information about the phenixbb mailing list