Dear colleagues,
here some comments about our Hg ripple problem - many thanks to all your
answers:
Indeed we were not able to completely get rid of the ripples (as stated
by Bart Hazes: not really affecting the biological conclusions - but
anyway!). Summarizing the suggested problems are: summation truncation,
anomalous scattering, anisotropy, multiple conformations, radiation
damage, occupancy.
We already did the occupancy refinement beforehand, which helped as well
as the anisotropic B refinement. Changing resolution limits did not help
much - anyhow unfortunately we do not have higher resolution collected
and data stops at the overall I/sigI at 3 sigma. Changing from Hg to
Hg+1 (compound originally was ClHgCH3) helped in decreasing R-factors,
although I always thought there should be covalent binding to Cys.
Changing f' did not seem to have a big impact, although it is clearly
different (approx. -11) from the standard CCP4 library for CuKa (-5.0)?!
BTW: the Hg peaks are only 25 sigmas, and not 80 and dual conformations
were not obvious.
Thanks to: ClemensV, KayD, GeorgeS, BartH, DavidS, KevinJ, EleanorD,
TassosP, TommiK, PeterM
Klemens Wild
