CRYSOL: Maximum in I(q) at q=0?

Calculation of SAXS and SANS profiles (CRYSOL, CRYSON), superposition of models (SUPCOMB, DAMAVER, DAMCLUST), database DARA
Post Reply
Message
Author
dwinogra
Guest
Posts: 1
Joined: 2019.01.19 07:58

CRYSOL: Maximum in I(q) at q=0?

#1 Post by dwinogra » 2019.01.19 09:07

saxs_comparison.small.pdf
CRYSOL vs experimental result
(105.64 KiB) Downloaded 14 times
Dear ATSAS community,

I'm a computational researcher who wants to use CRSYOL to generate a SAXS profile from a PDB we have generated in our lab.
I want to compare my SAXS result (I(q) vs q) to a plot from experiment.
I do not wish to do the fit directly with CRYSOL, but rather to generate an independent profile from my PDB,
which I will then compare to data from experiment.
I have three questions relating to this, and any help would be greatly appreciated.

Using CRYSOL Version 2.8.3, I was able to generate a profile, which appears similar in overall shape to the one from experiment.
However, the experimental curve has a local minimum in I(q) at q=0, whereas my I(q) curve generated from CRYSOL has a maximum at q=0.
(1) My first question is, why is this the case?

The experimental paper includes the sentence:
"Radial averaging, intensity scaling, and background subtraction were performed by MarParse."
So, one possibility is that the experimentalists have a local minimum in I(q) where q=0 because they subtracted out the background,
which would have lead to a maximum at q=0.

(2) If that is indeed the case, is there an option for running CRSYOL where I could eliminate the background,
and then my I(q) profile would also have a local minimum at q=0?
From the manual, I saw CRYSOL has options for "Minimum relative background" and "Maximum relative background",
but I do not know how to use/modify them.
Could modifying these options essential perform background subtraction?

I should also mention that my PDB structure does not include any water or ions.
(3) Could the addition of water and ions make a difference in the I(q) vs q profile, shifting I(q) down at q=0?

Lastly, I've read through the CRYSOL manual, and, if possible, could you suggest a resource with more information on how CRYSOL works (in addition to the 1995 paper), and more examples of its use (a tutorial perhaps, if there is one).

Thanks,
David Winogradoff

Attached I've included:
the logfile from my CRYSOL run "all-atom-old12401.log", and
the graphed output "saxs_comparison.pdf" (red from CRYSOL, blue from experiment).
saxs_comparison.small.pdf
CRYSOL vs experimental result
(105.64 KiB) Downloaded 14 times
Attachments
all-atom-old12401.log
CRYSOL log file
(2.48 KiB) Downloaded 10 times

User avatar
AL
Administrator
Posts: 670
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Guinier law

#2 Post by AL » 2019.02.05 11:31

dwinogra wrote:
2019.01.19 09:07
...
However, the experimental curve has a local minimum in I(q) at q=0, whereas my I(q) curve generated from CRYSOL has a maximum at q=0.
(1) My first question is, why is this the case?
Hard to tell. Collecting scattering data close to q=0 is challenging - that's where the primary beam hits the beam stop. Experimental data are typically collected up to some small qmin; to obtain the I(0) value (and the radius of gyration) the Guinier approximation is used.

Assuming your experimental data are from a homogeneous monodisperse solution I'd just remove the data close to q=0 that doesn't follow the Guinier law.

Post Reply