Positive difference map features indicate a likelihood that atoms
should be placed in that location. Putting atoms in that density, any
atoms, will lower the R value. This does not mean that your
interpretation is correct.
The fact that you, the person who has seen this map most clearly,
can't decide between a PO4 and a Ade means that this density is
ambiguous. These are two very different shapes!
A constellation of isolated blobs is about the worst thing to figure
out. It is quite possible that you have a partially occupied something
with ordered water molecules when that thing is not present. When
looking for partially occupied things you have to contour the map at a
much lower level, but not take what you see too seriously. Remember you
are looking for something at, say, 50% occupancy, with 50% occupied
water molecules sitting on top.
That said, I don't understand you difference maps. You built a ADN
into your blobs, and that molecule is quite a bit larger than your
blobs, yet you have a huge amount of positive difference density
covering your ADN. How is this possible? The goal of refinement is to
make the difference map go to zero at the location of atoms. There is
something about these maps that you are not telling us.
When interpreting blobs the first and most important thing to
consider is what is in the crystal. If it doesn't contain PO4 there is
no reason to test PO4 as a possibility. If you didn't add ADN then you
are hypothesizing that it was carried all the way through purification
which means that has to bind fairly strongly to survive even at partial
occupancy. Is this location on your protein a nucleotide binding site?
These things are easily recognized just from the structure of the protein.
Does the hydrogen bonding and charge-charge interactions of your
model make sense? It is hard to tell in a flat picture, but I don't see
many hydrogen bonds to your ADN model. If I compare your PO4 model to
the structure I see in the ADN map, I see that there is a ASP right next
to that blob. You can't put a PO4 next to an ASP.
Since the map is confusing you have to use as much information from
it as possible (lower contours) but add in as much information from
other sources as possible. What's in the crystal? What is your
cryoprotectant? Is this part the the protein a known binding motif for
something? What is the function of this protein and what sorts of
compounds might be expected to bind to it? Is this an enzyme and is the
spot anywhere near the active site? Do you know where the active site is?
Once you build a model you have to test it. You should be your worst
critic! As I said, a drop in R value is meaningless. Does the
chemistry make sense? Can you explain why that molecule is there? Does
it have a purpose? Can you perform an experiment that confirms your
model? Can you soak more of that compound into the crystal and see an
increase in occupancy? Can you analyze the crystal by some other means
to detect the compound? Mass Spec? If PO4, can you detect the presence
of Phosphorus? The validation has to be designed based on your model.
You can get better maps than simple Fo-Fc maps. I really like the
maps produced by Buster and have great success with getting better views
that way. It is possible that the "Polder" map from Phenix might help -
I'm not too clear on the difference between it and the Buster map.
Load as much information into your head as you can, get the best
possible map(s), and start running though the possibilities! And be
willing to accept that you may never figure it out. Don't build a model
you don't believe.
I once spent about twenty years trying to figure out a blob (not full
time!). I got a nice paper about it in the end.
On 8/5/2018 7:00 AM, Preeti Preeti wrote:
> Dear CCP4 member
> I am solving a protein structure with Resolution 2.2 Angstrom. I could
> see some blobs (2fo-fc @1sigma) and fo-fc @ 2.5 sigma) and need your
> suggestions on these extra electron densities. In addition to this, in
> one of the large blob I have added the phosphate group.
> This is a nucelotide binding protein however structure I am showing
> here was of native protein crystal without any nucleotide soaking and
> also no phosphate buffer was used at any time during the purification as
> well as crystallization process.
> Furthermore in one such blob I tried fitting the adenine moiety though
> it is not fitting exactly in the map, it decreased the Rfree value
> kindly suggest me what it should be corresponding to?
> Also please let me know if any other structure information required
> regarding this protein data or this blob density
> Thanks a lot in advance
> To unsubscribe from the CCP4BB list, click the following link:
To unsubscribe from the CCP4BB list, click the following link: