How to normalize/process I(0)

Interactive and automated data processing tools (PRIMUS, GNOM, AUTORG).
Scattering from simple bodies (BODIES), peak analysis (PEAK), data plotting (SASPLOT) etc.
Post Reply
Message
Author
natskawinska
Guest
Posts: 1
Joined: 2019.08.07 09:46

How to normalize/process I(0)

#1 Post by natskawinska » 2019.10.10 11:23

Hello everyone,
I am working with several SEC-SAXS curves obtained at Petra III during different trips. I want to estimate the molecular weight of my protein using several different methods, including the equation MW = (NA*I(0)/c) / ΔρM2, where I(0) is the extrapolated intensity at q=0.
The intensities from different trips differ by a factor of 10, despite the same protein concentration. The result is that I get MW estimates 1000x too high, or 10000x too high, depending on the date of the data collection. I have checked all the units etc and there is no mistake in my calculations. When I let PRIMUS do the calculations, I get values in the correct (and expected) range. It seems that, in addition to protein concentration, my intensities need to be normalized by some parameter that is present in the files. Can anybody tell me which parameter that is?

User avatar
AL
Administrator
Posts: 695
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Re: intensities need to be normalized by some parameter

#2 Post by AL » 2019.12.03 12:48

natskawinska wrote:
2019.10.10 11:23
I am working with several SEC-SAXS curves obtained at Petra III during different trips...
If you want to use I(0) from the Guinier approximation for MW calculations you need data collected on an absolute scale and you need to know the concentration in each frame. In theory, you can use the UV trace (absorption at 280 nm) to calculate the concentration. At P12 the UV trace is not stored in the .dat files but is stored in a separate file (if it was recorded during the experiment). I(0) from SEC-SAXS measurements is rarely used for MW estimation in practice. Please use MW estimation methods which are not relying on the absolute scale I(q).

Post Reply