On 9/19/2014 2:29 PM, Eleanor Dodson wrote:
> hmm - crystallographically difficult. The usual way is to make a
> dictionary file from the chemical information about the ligand. try to
> build something obeying the chemical restraints into the density -
> refine those coordinates and validate them.
> Eleanor
>

   I would be a lot more cautious than this. It is good to remember that
the density for a fully occupied, ordered ligand will be as strong and
clear as that of the protein in the active site. The fact that you have
done some refinement but are still not sure if your ligand, or anything
for that matter, has bound means that at best your crystal has problems.

   You have two questions. Did anything bind during your soak? Did the
ligand of interest bind? The first you can attack fairly unambiguously
with an Fo(holo)-Fo(apo) map. I presume you have an isomorphous apo
data set since you are performing a soaking experiment. A clear set of
density that stands out above the background in the Fo-Fo map tells you
that something new is in your crystal.

   Figuring out what it is is another matter and is very difficult if
the density is weak. Eleanor is recommending building and validating.
Unfortunately none of the validation tools we have have hard cutoffs of
what is acceptable and what is not. With years of experience one can
come to a conclusion, but someone starting out doesn't really know if
those tests are definitive. Remember you are not trying to decide if
your ligand bound, but if the thing that bound is your ligand. There
are many other things that could happen.

   Maybe your ligand bound. Maybe something else in your solution
bound. Maybe the chemical in your solution wasn't what you thought it
was. Maybe it was, but has been chemically transformed since then and
is now something else. Maybe this particular crystal had something in
it all along that your apo crystal didn't.

   This isn't a matter of seeing IF you can build your ligand into the
density and get away with it. You have to also say that you CAN'T build
anything (reasonable) in that density OTHER THAN your ligand.
With weak and fuzzy density this is very hard to do.

   If you performed a soak and got something to bind, maybe you could
soak longer or at higher concentration and get a more definitive map.

   If you have not solved the structure of a protein:ligand complex
before I suggest you check out other models in the PDB with good, strong
density and see what a full power ligand looks like in a map. Only
settle for weak density if there is no alternative and never settle for
ambiguous density.

Dale Tronrud

>
> On 19 September 2014 22:06, ansuman biswas <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Dear All,
>
> I have collected a diffraction dataset from a crystal soaked in a
> solution containing the ligand of interest.
> After refining a few cycles, I can see some density in the active
> site pocket, but not so clear to model the ligand unambiguously.
>
> Is any tool available to validate whether the ligand is actually
> there or not ?
> The Twilight server appears to be for PDB files that have already
> been deposited.
>
> thanks and regards,
> Ansuman Biswas,
> dept. of Physics,
> Indian Institute of science
>
>