*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
There goes the summary [Boaz Shannan (BA), Mischa Machius (MM), Jan
Donhalek (JD) and Flip Hoedemaeker (FH) answered] :
- FH (and later one the lab itself here) pointed the solution: include one
more neighbour atom as alternate. That worked the way I want, and even
without the need for cloning so many atoms.
Taking advantage of the on-going question, how should this be considered
for long side chains ? I recall to have read a message in the ccp4bb, but I
cannot remember exactly, the might be some limitations to this, ie, you must
clone until at least a certain atom. I tried to recover this message, but I
could never set the right keyword to make it come out of google. Suppose you
have an alternate for just the NZ of LYS, where should the alternate go up
to ?
- Thanks to BS for recalling the convention from PDB. My resolution does not
allow to use shelxl, but should be enough to argue for an alternate
conformation.
- As MM remembers, it is worth to note that not only the (clearly) alternate
atoms have different position, but one ones to which they are attach might
many times be in slightly different places.
- JD points that many times there is the need to model the whole molecule
as alternate; maybe a water molecule might account for the "extra" (I must
say kind of weak) blob nearby. That shold be born in mind, but in this case
it looks more liek an alternate conformaiton for the oxgyen. Concerning
alternating the whole molecule, I observed again the same phenomenon
described below, similar molecules sharing (most of) the same position, in
spite of occ sum < = 1, have b-factors set apart and one pushed out of the
density.
Thanks for all that replied, I still raise the question about where to
clone the side chain atoms up to, including the refinement program
restrictions (geometrically, it seems sensible that if I clone a CB I should
also clone a CA) ?
Jorge
*** 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 a structure and I can see much "clearly" that a glycerol
is bound very near the active site with alternate conformation for one of
the oxygens (let's say, O3). I tried to "clone" sole this oxygen (geometry
and hydrogen bond contacts should allow for the two conformations, as the
electron density points) according to the rules for residues (A and B
flags - OK, I shoud still check how this is allowed in PDB rules for a
ligand), but refmac seems to ignore the geomterical restraints to these
"cloned oxygens" and the geometry is somehow distorted thereby. One
solution is to clone the whole glycerol molecule (what would increase the
number of parameters, cloning atoms that in principle would not need so),
but that goes again to the question that one of the alternates has its B
factor lowered while the other is increased during refinement. CNS is a
solution, but I will miss TLS in this case and yet the question of jumping
between programs with other alternate side chains, etc....
Any better solution for this ? Might I make a "fake" glycerol with two
alternate oxygens and generate a special dictionary for it ? Maybe this
question is not all irrelevant as cryocrystallography is around and each
time more high resolution data allow you to observe this.
Thanks,
Jorge
