Hi Tom, thanks for your reply!

Hi Matthias,

  B-factor is an atom attribute so I guess you are interested in average B-factor for a residue? 

Yes, per residue B-factor scores  would make sense to me – unless you are dealing with  models derived from very high resolution crystal structure data, experimental structures often anyhow only refine a per-residue-averaged B-factor…

For floppy surface side chains that could be large while the backbone atom B-factor is small, so maybe you also want average B-factor for backbone or side chain separately.  AlphaFold 2 and 3 save pLDDT as B-factor, ie per-atom, and AlphaFold 3 interestingly gives different pLDDT values on different atoms of a residue -- again maybe you would want an average.  For AlphaMissense there are 19 pathogenicity scores for the 19 possible mutations.  How do you envision plotting those per-residue?  An average doesn't seem too useful.  Maximum?  Show 20 separate graphs? 

Hi Tom, I think an average would be very useful. For example, I downloaded the alphafold missense data through ChimeraX and then used the mutationscores program to calculate a mean score:

#get the AF model
alphafold fetch Q86V81
#get the missense data
open Q86V81 fromDatabase alpha_missense format amiss
#associate missense data to structure
mutationscores structure #1
#label each residue with matric
mutationscores label #1 amiss height 3
#define a new score called avg by avergaing over all substitution scores
mutationscores define avg fromScore amiss setAttribute true combine mean
#color by avg missense score
color byattribute r:avg #!1 target csab palette 0.232289,blue:0.616121,white:0.999953,red
#scale cartoon by acg missense score
cartoon byattribute r:avg #!1 0.232289:0.25 0.999953:2

The resulting rendering for my example looks like this:

Comparing this to the missense heatmap from the Alphafold database suggests to me that averaging is a reasonable approach to identify hotspots where mutations are poorly tolerated (Note: we happen to know this protein fairly well, and these scores are an amazingly accurate predictor for regions that are involved in specific protein-protein contacts and functionally important. That’s why we are excited to generate these plots):

For B-factor and pLDDT I could imagine you might want graphs for every chain of a complex which could be painful to request 1 at a time for say a ribosome.

That is fair – one could argue that users are likely to look at one chain at a time or at isolated alphafold predictions, or hard-code this into how the command works, similar to how the `sequence chain` command does not spit out sequence viewers for the entire model when you supply a multi-chain model.

  The residue graphing should allow clicking or dragging on the graph to select corresponding residues.

That would be great to have indeed

  There a lot of unclear details about how this would work.

  In general ChimeraX tries to show residue attributes directly on the structure by colors (e.g. average bfactor, pLDDT, maximum mutation pathogenicity score).  I don't have a clear picture of when it is more useful to show those kinds of scores with a graph where the horizontal axis is residue number.  As Elaine points out, that makes more sense with a sequence alignment visualization.

I see the point. Is there an option to export attribute data that is associated to the structure by ChimeraX (such as conservation scores or the averaged AlphaMissense score) for plotting with external programs?

Many thanks,

Matthias

        Tom


> On Jan 29, 2025, at 9:27 AM, Elaine Meng via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
>
> Hi Matthias,
> Sorry, not in general.  Although you can see a histogram of attribute values in the Render by Attribute tool, the X-axis is the attribute value, not atom or residue number.
> <https://rbvi.ucsf.edu/chimerax/docs/user/tools/render.html>
>
> The Sequence Viewer can show various attributes as headers, but currently just things like sequence conservation and RMSD over associated structures.
> <https://rbvi.ucsf.edu/chimerax/docs/user/tools/sequenceviewer.html#headers>
>
> Best,
> Elaine
> -----
> Elaine C. Meng, Ph.D.                      
> UCSF Chimera(X) team
> Resource for Biocomputing, Visualization, and Informatics
> Department of Pharmaceutical Chemistry
> University of California, San Francisco
>
>
>> On Jan 29, 2025, at 2:44 AM, Vorländer,Matthias Kopano via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:
>>
>> Dear ChimeraX team,  I wondered if it is possible to generate a line or scatter plot of an attribute value (B factor, Alphafold pLDDT scores, or AlphafoldMissense score…) over the residue number?  For example something like this (taken fromhttps://journals.plos.org/plosone/article?id=10.1371/journal.pone.0089116) :
>>
>> <image001.png>Specifically, for the AlphafoldMissense data, such a plot would be more informative then the default histogram that is displayed when fetching the scores when looking for functional hotspots in your POI.
>> Many thanks,
>> MAtthias
>
>
> _______________________________________________
> ChimeraX-users mailing list -- chimerax-users@cgl.ucsf.edu
> To unsubscribe send an email to chimerax-users-leave@cgl.ucsf.edu
> Archives: https://mail.cgl.ucsf.edu/mailman/archives/list/chimerax-users@cgl.ucsf.edu/
>