[Chimera-users] the docking confidence score

[Elaine Meng]

Hi Yanhe,
As I understand it, there is no "magic" level of correlation that means the result is right or good across all systems -- instead it is for comparisons only within the same system, i.e. different fits of the same structure into the same map.  

There are too many variables (resolution, surface contour level, etc.) to compare between different systems.

I hope this helps,
Elaine
-----
Elaine C. Meng, Ph.D.                       
UCSF Chimera(X) team
Department of Pharmaceutical Chemistry
University of California, San Francisco

> On Jan 26, 2022, at 4:38 PM, Yanhe Zhao via Chimera-users <chimera-users at cgl.ucsf.edu> wrote:
> 
> Hello there,
> 
> Here is continuous question about sequential docking score:
> 
> By eyes I can find that models have good fitting to my EM, and the correlation is 0.8 (I calculate it in two ways: each or all of them). But I was curious in what range we could consider the docking reliable?
> 
> Thanks and cheers,
> Yanhe
> 
> Elaine Meng <meng at cgl.ucsf.edu> 于2022年1月25日周二 15:18写道：
> Hi Yanhe,
> Chimera does not calculate a single overall score for the sequential fitting of multiple structures.
> 
> Sequential fitting just fits one structure locally, then subtracts density from where that structure is, then fits the next structure locally, etc.  So structures are only fit (and scored) one at a time, and it depends what order they are in and also their starting positions since they are only fit locally.  However, you can keep iterating sequentially through all of the structures multiple times to improve the results.
> 
> See sequential fitting help
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/fitmap.html#sequence>
> ...and video
> <https://www.cgl.ucsf.edu/chimera/videodoc/FitSeq/>
> 
> So you have to use your own knowledge of the system both in initially placing the structures (although you could do that by a previous fitting calculation) and in evaluating the results.
> 
> To get correlations for each of the structures in their current positions after sequential fitting, you can use the Fit in Map tool (use the Options to turn off rotation and shift if you don't want them to move any more) or "measure correlation" command (which will give numbers in the Reply Log opened from Favorites menu).
> 
> Maybe you could make a single molmap from all of the structures together and then use "measure correlation" to compare it with the original experimental map to get an overall correlation.
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/measure.html#correlation>   
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/molmap.html>
> 
> Another way to evaluate the results is to look for clashes between the atomic structures (e.g. Find Clashes tool or findclash command). 
> <https://www.rbvi.ucsf.edu/chimera/docs/UsersGuide/midas/findclash.html>
> 
> I hope this helps,
> Elaine
> -----
> Elaine C. Meng, Ph.D.                       
> UCSF Chimera(X) team
> Department of Pharmaceutical Chemistry
> University of California, San Francisco
> 
> 
> > On Jan 25, 2022, at 12:23 PM, Yanhe Zhao via Chimera-users <chimera-users at cgl.ucsf.edu> wrote:
> > 
> > Greeting there,
> > 
> > I'm doing the sequential fitting to fit 3 models to my map. The question is where I can find the docking confidence score?
> > 
> > I did try molmap to simulate the map of one model and then use the  "Fit in map" to have the correlation readout. 
> > But I'm not sure is the correlation good to describe confidence, and what I want is the score to describe overall confidence of sequential fitting not just one model.
> > So any log file I can find to report the score of docking?
> > 
> > Thanks and cheers,
> > Yanhe
> 
> _______________________________________________
> Chimera-users mailing list: Chimera-users at cgl.ucsf.edu
> Manage subscription: https://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users