[phenixbb] Filtering CCP4 maps to lower resolution

Benjamin Barad ben.barad at ucsf.edu
Fri Nov 14 10:16:09 PST 2014

> complete representation of your map (the one in CCP4 formatted file) is a
> reciprocal space box of reflections with miller indices being |h|<N1/2,
> |k|<N2/2, |l|<N3/2 , where N1,N2,N3 is map gridding.
> Converting map (from CCP4 file) into a set of structure factors in a
> sphere of given resolution is an operation associated with loss of
> information. Further truncation of resolution obviously results in further
> loss of information. All in all this means the artifacts you observe after
> these manipulations are expected.

My understanding is that the loss of information would happen at high
resolution (which I want to be filtering out anyway), so it would result in
smoothing rather than generation of lots of noise that I would associate
with loss of low-resolution information. Am I incorrect in that

If so, is there a better way to do a low-pass filter on real-space data in
phenix? One that would only remove high-resolution information?

> Alternatively, I guess, you can do local averaging: replace each map grid
> node with the average takes over all 27 nearest neighbors of that node or
> all neighbors in a sphere of R around it (Wang's averaging). As A. Leslie
> demonstrated original Wang's averaging can be done efficiently in
> reciprocal space by convolution of structure factors with some function
> (see that paper for details).

> I can write a simple Python script for you to do this task.
> Pavel

This is an interesting option, and I would like to try it... Working in
reciprocal space would require an FFT, however, which you suggest above
would result in the same loss of data, right?

Thanks very much,

> On 11/13/14 4:00 PM, Barad, Ben wrote:
> Hi All,
> I am currently trying to filter a CCP4-formatted real space map from the
> EMDB, which is at 3.2 Å resolution, to different resolutions from 3.2 to 5
> Å, and I've been having a bit of a difficult time getting it to work.
> My first thought was to do an fft to get structure factors, then
> regenerate the map from those structure factors with a resolution cutoff.
> To do this, I used the tools "phenix.map_to_structure_factors" and
> "phenix.mtz2map" to generate a structure factor file, then convert it back
> into a map. When I do this, I end up with a very noisy map that is not
> usable for what I would like to do.
> Is there a better way to accomplish this in phenix (or another tool)? I've
> also tried some low-pass filtering with proc3d, without much success...
> Thanks,
> Ben Barad
> Fraser Lab @ UCSF
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20141114/006e3dea/attachment.htm>

More information about the phenixbb mailing list