*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Prem,
If you are seeing non-protein peaks at 3.5A, I am willing to bet they
are ions.
If one calculates an electron density map from any PDB file with some
waters in it and roll off the resolution to ~2.9A, you will find that
the water atoms are not visible when you contour the map at "1 sigma".
I made a movie depicting this phenomenon years ago:
http://bl831.als.lbl.gov/~jamesh/movies/resolution.mpeg
Each frame of the movie "worsens" the resolution. You can see the waters
vanish at 2.9A. They are still there, of course, and if you contour
lower you can see them. But you just don't see them at 1 sigma.
Remember, this is a calculated map, so the "1 sigma" value is dominated
by the RMS density of protein. I think this is more or less true in all
but the noisiest 2Fo-Fc maps. Gerard probably knows better.
Now I'm sure I will hear some comments on how "portable" the MPEG-1 file
format is....
-James Holton
MAD Scientist
Bart Hazes wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
I don't recall if "thou shalt not place waters at 3.5A" is among the
10 commandments for protein crystallographers but doing it will
certainly raise eyebrows.
In that resolution range you are in observation/parameter limbo and
each water adds 3 (or 4) more parameters and, unlike protein atoms,
there are no real restraints on those parameters. In any map you are
going to have noise peaks and some of those will be in positions that
could make sense as a water. At this resolution you just can't tell if
you are modeling noise or reality and I think you are better off not
modeling waters. If you have good evidence that there should be a
water, for instance a catalytic site water that has been seen
consistently in related structures at higher resolution, then I would
look if the density would support that water to be present but still
be rather cautious in the way to describe it in a paper.
Bart
Prem kumar wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear all,
I am refining structure at 3.5 A resolution. the
protein has 340 residues. CNS water_pick gave me 80
water molecules. out of that ~50 are in resonably good
density. when i go through PDBs at this resolution
many people didnt add water. my question is, can I add
water molecules in this resolution, or upto what
resolution. please quote if any reference regarding
this case.
Thanks in advance!
with regards
prem
