Hard-to-believe changes in R-free

Ethan A Merritt merritt at u.washington.edu
Fri Jul 17 12:04:51 EST 1998

In article <35AF5A97.CF62A637 at aecom.yu.edu>,
Yoram Puius  <puius at aecom.yu.edu> wrote:
>SO, I did a wacky experiment:  I replaced every water with a B-factor
><20.0 with a sodium with a B=20.0, and did 30 cycles positional refine-
>ment in X-PLOR 3.851 and 20 cycles B-refinement.  (I used a bulk solvent
>correction for data up to 20.0A)
>Here's what I got:
>Water   Refined H2O B   r               r_free          Refined Na+ B
>-                       0.184083        0.248373        -
>4       13.43           0.183829        0.246176        23.78
>                                        ^^^^^^^^
>Note the 0.2197% drop in R-free when water 4 is changed to Na+.

>Aside from it being a coincidence, the only other explanation I
>could come up with was that, since Na+ allowed closer non-bonded
>contacts, this actually allowed a generalized relaxation of
>all nearby coordinating atoms -- i.e. geometry and electron density
>no longer had to fight each other so much because you could squeeze
>the Na+ closer to the oxygens.

If you want to test this explanation you could run the refinement
again with the protein atoms fixed. You could even run it twice:
1) as before with the protein fixed,
2) starting from your model 4 above, changing the Na back to OH2 
   and again fixing the protein coords.

As you say, it would be better to back up your interpretation with
chemical evidence rather than simply pointing to R-free in any case.

					Ethan A Merritt
					merritt at u.washington.edu

More information about the Xtal-log mailing list