Molecular weight estimation

[EduardoBruch]

Hello,
For my usual data proccesing I was allways using the BSA standard or the Porod Volume/1.6 to estimate MW.
I just started to used the new version of Primus (from 2.8.3 package) and I found the new Molecular weight analysis option very useful.
I was wondering if there was any help/manual with further details about the different results, specially regarding the approaches used for each of the four automatic calculations (Vc, Qp, MoW and size and shape).
Thanks in advance

Eduardo

[AL]

I just started to used the new version of Primus (from 2.8.3 package)...

If your OS is Windows or Linux - please note that the latest package is 2.8.4. (If you have a Mac please stick to 2.8.3 for now.)

I was wondering if there was any help/manual with further details about the different results, specially regarding the approaches used for each of the four automatic calculations (Vc, Qp, MoW and size and shape).

Check out Hajizadeh et al. (2018) Consensus Bayesian assessment of protein molecular mass from solution X-ray scattering data Sci. Rep. 8:7204 doi

[EduardoBruch]

That is exactly what I needed, about the version I am using the 2.8.3 in mac.
Thanks Al!

[EduardoBruch]

Hello
I wanted to share a curiosity I found with some of my data. I am processing the saxs data from 2 homologous proteins from 2 different organisms with one 20% smaller than the other. The data looks good for both and surprisingly the Bayesian molecular weight is not only similar but exactly the same, more precisely 318450 Da (althought Rg, Dmax are similar but not the same). At the beggining i thought there was a problem with the data but seems to be ok. I was wondering if there might be a reason for this or is just a big coincidence.
Thanks in advance
See you

Eduardo

[franke]

The data looks good for both and surprisingly the Bayesian molecular weight is not only similar but exactly the same, more precisely 318450 Da (althought Rg, Dmax are similar but not the same).

You didn't quote the MW range and probability for it, but I guess they are at least overlapping. The source data, i.e. the PDB is somewhat sparse for entries >>100kDa. Hence we had to increase the bin sizes to accommodate for that and not to end up with bins with one or even no entry. You can see this in Figure 3 of the article linked above. Panel A shows the linear distribution of the molecular weight within the PDB. The '5' on the x axis indicates 500kDa, aim slightly right of the middle of 0 and 5 and you see how many entries there are - not many. Panel B is somewhat better; take log(318000) ~= 12.7 and you see that is the tail region of anything that is this size or larger (up to many MDa).

What would be interesting in your case: what are the ranges quotes for the two data sets (and what's their probability), and do the expected values fall into these ranges?

[EduardoBruch]

Hello all,
I have a new question about MW estimation. I am calculating the MW using the tool from primus and I found a bigger than usual difference in the estimation by the different techniques.
real MW= 750 Kda, calculate using primus QP=700 Kda, MoW 420 Kda, Vc 442 Kda, size and shape 660 kDa and finally the bayesian gives MW estimate 480Kda (with 56% MW probability) and a range between 372-505 Kda
The protein is an elongated and its properly folded but it has a flexible region in the extremes (this was verified by EM) The kratky plot shows a bell shape with a slow decay followed by an increase which is coherent with this. I was wondering if this flexibility could justify this MW estimation variance and if giveng the propertis of the protein any of the values should be more reliable.
Thanks in advance for your help!!
Kind regards

Eduardo
PS: I included this question in this post so I dont generate a new post.