Error during MW analysis

[Lewis Owen]

Hi, 

I am receiving the below error when trying to analyse the MW of my samples. It is not occurring for all samples. I've attached two example files, one that it works with (NND-2_pure) and one it doesn't (NND-2_Zinc). This is my first time analysing SAX data, so I suspect I have done something wrong during the Guinier Fit process. 
Error Message:

An unexpected error has occurred, please report it to the developers.
System: Windows-10-10.0.26100-SP0
RAW version: 2.3.1
Prebuilt: True
ATSAS version: 3.2.1

Error:
Traceback (most recent call last):

  File "bioxtasraw\RAW.py", line 16240, in run_with_except_hook

  File "threading.py", line 982, in run

  File "bioxtasraw\RAWAnalysis.py", line 3255, in _calcMWThread

  File "bioxtasraw\RAWAnalysis.py", line 3353, in calcVcMW

  File "bioxtasraw\SASCalc.py", line 1009, in calcVcMW

IndexError: index 1 is out of bounds for axis 0 with size 1

[Jesse Hopkins]

Hi Lewis,

You’ve found a bug in the Vc molecular weight calculation. Essentially, RAW has to make some decisions about where to cut off the profile in q to do the calculation and your input profile is breaking the algorithm for those decisions due to a bug. I appreciate your bringing it to my attention.

There’s a couple of things you can do to work around the bug. If your q units are in 1/Angstrom, then manually assigning the profile unit will help. You can do this by right clicking on the profile in RAW, going to Other Operations->Set q units and selecting 1/A. If the q units are in 1/nm, then you can first rescale the data into 1/A by going to the same Other Operations->Convert q-scale-> q/10, then manually set the units to 1/A as described above. Note that if you convert the q scale you’ll need to redo the Guinier fit before you calculate the MW.

I would be remiss though if I didn’t tell you that this data has some serious issues. First, your scattering profiles go negative at high q, which shouldn’t happen. That’s indicative of a significant buffer subtraction issue. Second, assuming that this data is a conventional macromolecular sample that you expect to be monodisperse in solution (the assumption you need to make for a good Guinier fit, but not sure if that’s what you’re expecting here), you’re seeing extremely strong signs of aggregation in solution, which means the data is not useable, even if you had a good buffer subtraction. While it’s possible that this is data of something else, in which case you can ignore this, if you’re trying to learn about the shape of a macromolecule in solution you need to fix your sample and buffer prep and redo your data collection, this data will not give you reliable results.

I’d suggest that you read this:

https://bioxtas-raw.readthedocs.io/en/latest/saxs/saxs_guinier.html

And listen to my lecture about basic SAXS data validation and analysis:

https://www.youtube.com/watch?v=8pgdRiEDGBQ

Which should give you a better idea of how to identify problems with your data, and what issues those might cause with the analysis.

If you’re interested in learning more about SAXS in general, please check out the resources in the Tutorials section of my beamline’s website:

https://www.bio.aps.anl.gov/pages/how-to.html

If you’re interested in recollecting the data I could also help with that (if you’re in the U.S., since that’s where my beamline is).

All the best.

- Jesse

----
Jesse Hopkins, PhD
Director
BioCAT, Sector 18
Advanced Photon Source

--
You received this message because you are subscribed to the Google Groups "BioXTAS RAW" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bioxtas_raw...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/bioxtas_raw/1c93f851-95fe-4a1d-bd73-25a512e7d92bn%40googlegroups.com.

[Lewis Owen]

Hi Jesse, 

Thank you for all of that information, I will definitely look into those tutorials!

My samples are small proteins (the ‘pure’ sample) and then small proteins that have been incubated in a solution of metal, in this case it was Zinc. This is because I found evidence in other data that the metals were causing the proteins to dimerise. When I did my background subtractions I subtracted the relevant buffers from the samples, e.g. water from the pure protein and the  zinc sulphate buffer from the protein+ zinc sulphate sample.

I assume that none of this info changes  your advice regarding the issues with the data? 

I’m not surprised to hear that there are problems with the data, after following the Bioxtas Raw tutorial with the tutorial data I noticed that mine was behaving very differently but wasn’t sure why. Thank you for explaining what the problem is! Unfortunately I’m based in Australia and unable to get time on the beamline again, but this data was just a value add to a larger project so if it’s not usable that’s ok. I really appreciate you bringing that to my attention as it will save me a lot of time trying to make sense of data that will never produce a usable result. 

Kind regards, 

Lewis 

---

From: Jesse Hopkins <jesse.b...@gmail.com>
Sent: Sunday, October 12, 2025 2:21 am
To: bioxt...@googlegroups.com <bioxt...@googlegroups.com>
Cc: lewis...@gmail.com <lewis...@gmail.com>
Subject: Re: [Bioxtas Raw] Error during MW analysis

 

[Jesse Hopkins]

Hi Lewis,

Your samples definitely aren’t monodisperse, or even a combination of monomer and dimer like you might hope for in the zinc buffer. In that case I don’t think you should try to use the data and wouldn’t recommend spending a lot of time on continued analysis.

SAXS is quite sensitive to sample preparation, and to get good SAXS data requires the sample be homogenous and monodisperse in solution. Your sample looks like it has a lot of something larger in solution (likely large aggregates). This could either be due to your solution conditions, or a lack of proper sample purity. If you were interested in pursuing the project I’d recommend listening to this lecture about sample prep and biophysical characterization to do prior to the experiments:

https://www.youtube.com/watch?v=uWonjUMrKI8

You’d probably need to improve your initial purification of the samples (e.g. do a size exclusion step and take peak fractions that are pure), redesign your buffer conditions (most samples aren’t happy in just water, you need at the very least some kind of buffer with defined pH and a bit of salt), do a better buffer match (you need to prepare buffer matches by dialysis, not just mixing by hand), and do at least some preliminary characterization (is the sample single peak on an SEC, do you get a single band on a gel are good places to start) to ensure it’s good enough quality for the experiment.

Hope that helps.

All the best.

- Jesse

----
Jesse Hopkins, PhD
Director
BioCAT, Sector 18
Advanced Photon Source