EOM ensemble has poor fit to scattering at high q

Linear (OLIGOMER), and non-linear (MIXTURE) analysis, singular value decomposition (SVDPLOT), addition of missing fragments (BUNCH, CORAL), analysis of flexible systems (EOM/RANCH & GAJOE), flexible refinement of high-resolution models (SREFLEX)
Post Reply
Message
Author
Rhyswg
Guest
Posts: 2
Joined: 2018.09.12 08:27

EOM ensemble has poor fit to scattering at high q

#1 Post by Rhyswg » 2018.09.12 08:39

Hi All,

I'm performing some EOM to fit the scattering curve of a protein I know consists of two flexible domains. The fit is okay at low q, but at higher q it diverges and actually crosses 0I at 0.25 1/ang (see attached). This doesn't seem credible to me and so I wonder if I'm doing a poor job of parameterising the run, or there is some software or data problem.

Any ideas?

Cheers,

Rhys
Attachments
profiles_001_1.png
profiles_001_1.png (7.28 KiB) Viewed 150 times

User avatar
AL
Administrator
Posts: 658
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

gajoe --no-constant

#2 Post by AL » 2018.09.17 13:29

Rhyswg wrote:
2018.09.12 08:39
The fit is okay at low q...
It's not okay - the Guinier region is not fitted: the Rg of your experimental data is clearly greater than Rg of your fit.
Does the molecular weight estimated from your experimental data match the molecular weight calculated from your protein sequence?
Rhyswg wrote:
2018.09.12 08:39
... but at higher q it diverges and actually crosses 0I at 0.25 1/ang (see attached).
Try re-running gajoe with '--no-constant'. See EOM manual for details.

Rhyswg
Guest
Posts: 2
Joined: 2018.09.12 08:27

Re: EOM ensemble has poor fit to scattering at high q

#3 Post by Rhyswg » 2018.09.18 11:39

Yea the MW from the experimental scattering is a pretty good match for the sequence (113 kDa to 105 kDa).

The CRYSOL fit for the crystal structure is much better than what EOM managed. Not well, but much better. I know the proteins consists of two flexible domains and I know where the linker is. Moreover, the CRYSOL fit for the individual models output from EOM fit much better than the curve I posted. This is a reasonably uninformed perspective, but it seems like the model selection part of the process is failing somewhere. Any ideas?

I performed EOM on a homologous protein I have SAXS and X-ray scattering for and it performed very well.

Cheers,

Rhys

Post Reply