[ccp4bb]: Riding Hydrogens

[Jason Yano <jkyano@scripps.edu>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Hi All,

Another subject for debate. I am in the process of refining a 2.1 angstrom
structure using the latest version of Refmac. The default for restrained
refinement seems to be to generate all hydrogens. This doesn't seem very
resonable at 2 angstrom resolution. I usually change the default to not
include the hydrogens, but I was a bit sleepy one afternoon and forgot to
change the default. Upon noticing that I had set it wrong, I decided to
re-run refinement again. Surprisingly, I found that adding the hydrogens
in lowered both R and R free by about 1%. Still skeptical, I tried a 2.3
angstrom data set, in and early stage of model building, here are the
results:

without hydrogens:

REMARK   3  FIT TO DATA USED IN REFINEMENT.
REMARK   3   CROSS-VALIDATION METHOD          : THROUGHOUT
REMARK   3   FREE R VALUE TEST SET SELECTION  : RANDOM
REMARK   3   R VALUE     (WORKING + TEST SET) : 0.23753
REMARK   3   R VALUE            (WORKING SET) :  0.23305
REMARK   3   FREE R VALUE                     :  0.32008
REMARK   3   FREE R VALUE TEST SET SIZE   (%) :  5.2
REMARK   3   FREE R VALUE TEST SET COUNT      :  1196

with hydrogens:

REMARK   3  FIT TO DATA USED IN REFINEMENT.
REMARK   3   CROSS-VALIDATION METHOD          : THROUGHOUT
REMARK   3   FREE R VALUE TEST SET SELECTION  : RANDOM
REMARK   3   R VALUE     (WORKING + TEST SET) : 0.23091
REMARK   3   R VALUE            (WORKING SET) :  0.22655
REMARK   3   FREE R VALUE                     :  0.31141
REMARK   3   FREE R VALUE TEST SET SIZE   (%) :  5.2
REMARK   3   FREE R VALUE TEST SET COUNT      :  1196

So I am unclear on on the requirements for adding the hydrogens on in
refinement. At what resolutions do they become justified? Why are they
lowering my R and R-free?


Thanks,


Jason