There are lots of possibilities, but any model you build must match
the electron density of you map and chemical properties of your
proposal. If you have a small spherical blob it can be an individual
atom, perhaps decorated by hydrogen atoms.
Those possibilities with hydrogen atoms would be H2O, NH3, NH4+, H4C,
H2S, etc. You shouldn't be modeling in methane or hydrogen sulfide
unless you have some reason to believe these molecules are present in
your crystal (Your nose will tell you about H2S! Also the sulfur atom
has a much denser electron density cloud.) If you have NH4+ it will be
even closer to its hydrogen bond partners than water, so that won't help
you with distant blobs.
The other single-atom possibilities are metal ions, halogens, or
noble gases. As with methane, you should only build a noble gas into
your model if you worked hard to add that element to your crystal.
Metal ions are common, but like ammonium, their ligands will be closer
than those of water. Of the common halogens, chlorine and bromine will
have ligands slightly further away than those of water, but these ions
are very electron dense and unless they have partial occupancy will
really stand out.
Unfortunately, map artifacts due to series termination and errors in
phases will result in blobs of density floating through out the bulk
solvent region. These "features" tend to be stronger near the surface
of the molecule and fade as one moves out into the void but will build
up again on symmetry axes and special positions. Placing atoms in
"noise" peaks will just lock those errors into your model and prevent
you from achieving a good result in the end.
If I had a map with a blob 4 A from any atom in the model I would
just leave it empty. Your goal isn't to flatten the difference map, nor
to get the lowest R values, but to build the most likely model.
Dale Tronrud
On 6/20/2016 4:07 AM, Santhosh G wrote:
> After adding water molecules and successive refinement, still there are so many blobs resembling water but coordination is not 4 even distance is >3.2 A (in some cases).
>
> What might be possible atom/molecule that may occupy at these sites ?
>
>
> Please give your suggestions...
>
>
> Thank you
>
> Santhosh
>
|