[phenixbb] Fo-Fo difference map for non-isomorphous datasets

Jordan Luke Pederick jordan.pederick at adelaide.edu.au
Mon Mar 11 19:40:10 PDT 2019

Also, I am interested if anyone knows of any publications that have included a real-space difference map?



From: Jordan Luke Pederick
Sent: Monday, 11 March 2019 1:25:57 PM
To: Eric Montemayor; Pavel Afonine; Philip Kiser; Tim Gruene; phenixbb at phenix-online.org
Subject: Re: [phenixbb] Fo-Fo difference map for non-isomorphous datasets

Thank you all for the prompt responses and suggestions.

My metal ion of interest is Cesium in this case, so I guess I am more interested in using this approach for visualization rather than peak identification - the strong density of Cs is fairly obvious

I have attached a few images of the superimposed (Fo, Phi) maps and resulting difference map using either Superpose maps in phenix or transform_map in coot following Tim's suggestions.

I also tried using maprot (CCP4) to apply the rotation/translation operator from the LSQ superpose but the resulting map was not superposed correctly (not sure why).

Both results seem kind of reasonable - I have shown one of the difference peaks for what is likely to be a bound Cs. For both approaches the peak was very obvious at a 3 sigma level, and could still be observed to around 10 sigma in both cases. I am still unsure about differences in scaling, so although strong positive peaks can be identified, I wouldn't be confident using the map to interpret smaller differences.

In response to Eric - I checked the anomalous signal, which was fairly weak. The resulting anomalous difference map did show positive peaks up to a level of ~10 sigma that matched those observed in the above real-space difference maps, with a 3 sigma level corresponding to 0.0137 e/A^3. The CCanom was also low (1-5% between shells), so I was thinking that another approach might be more suitable for this data?



From: Eric Montemayor <montemayor.eric at gmail.com>
Sent: Sunday, 10 March 2019 4:50:21 AM
To: Jordan Luke Pederick
Subject: Re: [phenixbb] Fo-Fo difference map for non-isomorphous datasets

Why not just look for peaks in the anomalous difference Fourier for just the derivative dataset?  If there’s even a weak anomalous signal, you won’t need to mess with non-isomorphism between multiple datasets.... the information you need is already contained alone in the derivative dataset.

What is the metal?  The above approach won’t work for something like magnesium, but should work for almost any transition metal.


On Sat, Mar 9, 2019 at 3:23 AM Jordan Luke Pederick <jordan.pederick at adelaide.edu.au<mailto:jordan.pederick at adelaide.edu.au>> wrote:


I am trying to generate a Fo-Fo difference map for two data sets of the same protein at ~2.3 angstrom resolution. One dataset is native and the other was soaked with a metal ion, so the purpose of the map is for identifying and visualizing peaks that may correspond to a bound metal ion.

These datasets differ by 6 angstroms (~10%) along the A edge - an isomorphous difference map generated in Phenix GUI wasn't useful

As an alternative I was planning to make a real-space Fo-Fo difference map making use of Superpose Maps in Phenix Gui as follows:

Superpose maps would be used to superimpose (Fobs.metal, PHI.metal) and (Fobs.native, PHI.native)

The superimposed maps will be opened in Coot and a difference map generated using "Extensions -> Maps ->  Make a difference map"

There were two things I was unsure about:

Does this seem like a reasonable alternative?

Is scaling of the two datasets required before generating the difference map?



phenixbb mailing list
phenixbb at phenix-online.org<mailto:phenixbb at phenix-online.org>
Unsubscribe: phenixbb-leave at phenix-online.org<mailto:phenixbb-leave at phenix-online.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20190312/3a9a65a1/attachment-0001.htm>

More information about the phenixbb mailing list