SAXS MoW

Small angle scattering software (except ATSAS)
Post Reply
Message
Author
Matt
Member
Posts: 9
Joined: 2011.01.19 01:14
Location: Stanford, CA, US

SAXS MoW

#1 Post by Matt » 2011.07.07 03:50

Hi all,

does anyone also fail to use SAXS MoW in www.if.sc.usp.br/~saxs/ currently? Is the server down at the moment or is that my problem?
I am using Java 1.5.0_30 in Safari 5.0.5 in Mac OS X 10.5.8.

And is there any other way I can calculate the MW with my SAXS data without measuring the water scattering or scattering from a protein standard?

Thanks,
Matt

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

SAXS MoW alternatives

#2 Post by AL » 2011.07.07 11:34

SAXS MoW works for me (Java 1.6.0_26, Firefox 3 on Windows). The only thing that one has to keep in mind is that the data has to be in Å-1.
Matt wrote:And is there any other way I can calculate the MW with my SAXS data without measuring the water scattering or scattering from a protein standard?
Sure. You can use the simple command-line tool DATPOROD from the full ATSAS installation. It calculates the Porod volume from the "gnom.out" files. If you divide the Porod volume by ~1.7 you get the molecular mass.
Alternatively you can take the volume from a DAMMIF (or DAMMIN) model and divide it by 2 to estimate the molecular mass.
If the experimental data is in nm-1 the volume will be in nm3 and the molecular weight in kDa; if the data is in Ǻ-1 the volume will be in Ǻ3 and the MW in Da.

Matt
Member
Posts: 9
Joined: 2011.01.19 01:14
Location: Stanford, CA, US

Re: SAXS MoW alternatives

#3 Post by Matt » 2011.07.09 09:35

AL wrote:SAXS MoW works for me (Java 1.6.0_26, Firefox 3 on Windows). The only thing that one has to keep in mind is that the data has to be in Ǻ<sup>-1</sup>.
Thanks AL, yes, it works for me in Java 1.6 IE on Windows as well, just can't get it worked in Mac....the MWs that I got from SAXS MoW is not quite matched my expected MWs and the estimated MWs using I(0) by guinier analysis with water scattering though. And here is what I got using I(0) by guinier analysis with water scattering:

1st protein:
mg/mL / Rg / MW
0.3125 / 46.16 / 167.53
0.625 / 46.04 / 182.97
1.25 / 45.21 / 176.42
2.5 / 46.25 / 186.42
5 / 47.03 / 191.83

2nd protein:
mg/mL / Rg / MW
0.1874 / 51.00 / 164.99
0.375 / 51.04 / 167.87
0.75 / 54.46 / 178.4
1.5 / 53.36 / 165.08
3 / 53.62 / 170.18

Note that Rgs are from Autorg and the MWs are calculated using an in-house software at the beamline.

My question is, do you think the increases of the Rg/Mw with concentration are severe enough that indicating aggregation of proteins?

Many thanks again,
Matt

james2009
Active member
Posts: 30
Joined: 2009.02.03 16:56
Location: University of Sydney

Mr

#4 Post by james2009 » 2011.07.14 12:55

What's the expected Mr for both proteins?

For protein 1 there are signs of "either" oligomerisation i.e. not necessarily aggregation, but again it could be a concentration error for I(0) analysis.

For protein 2, I would say you're within experimental error, i.e. determination of concentration

You should try to study protein 1 further, by MALLS or preferably a complete in solution method like AUC (either sedimentation equibilibrium or velocity) to provide you with futher details as you how your system operates.

Cheers, James

james2009
Active member
Posts: 30
Joined: 2009.02.03 16:56
Location: University of Sydney

#5 Post by james2009 » 2011.07.14 12:56

p.s. Guinier region will help tell you if there's aggregation

james2009
Active member
Posts: 30
Joined: 2009.02.03 16:56
Location: University of Sydney

#6 Post by james2009 » 2011.07.14 12:57

p.p.s. it also assumes there are no higher MW contaminants effecting the answer

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

#7 Post by AL » 2011.07.15 13:49

Matt wrote:My question is, do you think the increases of the Rg/Mw with concentration are severe enough that indicating aggregation of proteins?
Not really. The Rg changes are small (±2-3%) and do not follow a specific trend. Have a good look at the curves - do they overlap well?

Matt
Member
Posts: 9
Joined: 2011.01.19 01:14
Location: Stanford, CA, US

#8 Post by Matt » 2011.07.24 02:46

AL wrote:
Matt wrote:My question is, do you think the increases of the Rg/Mw with concentration are severe enough that indicating aggregation of proteins?
Not really. The Rg changes are small (±2-3%) and do not follow a specific trend. Have a good look at the curves - do they overlap well?
Thanks for the replies, James and AL.
Actually the reviewers of our submitted paper want to know the stoichiometry of both complexes from the SAXS I(0).

Both of them look good in the Guinier region, so I also don't think there is sign of aggregation.

The expected size of the 1s protein is 178 kDa, the 2nd one is 196 kDa. The 2nd one is a Fab-protein complex, my protein is ~146 kDa and I assume the Fab is about 50 kDa (we don't have the sequence of the Fab though). I have a every good average ab initio envelope (NSD = 0.83 +/- 0.057) for the Fab-complex and I can easily fit my protein and the Fab in it, the excluded volume is 4.663e+5 Ǻ3.

I have re-processed my data using manually determined Rg values (autoRg frequently chooses higher q range (sRg>1.3)...), I satisfy with all the numbers for the 1st one, but not the 2nd one, so I just post the data of the 2nd one:

mg/mL / I(0): guinier / I(0)/C: guinier / I(0)/C: exp / Rg / MW
0.1875 / 571.31 / 3047 / 3290.1 / 53.192 / 169.36
0.375 / 1126.4 / 3003.7 / 3568.8 / 50.932 / 166.95
0.75 / 2415.5 / 3220.7 / 3557.2 / 54.028 / 179.01
1.5 / 4449.2 / 2966.1 / 3186.4 / 52.642 / 164.86
3 / 9185.6 / 3061.9 / 2969.1 / 53.734 / 170.18

From the numbers, I believe both are 1:1 complexes. But I don't understand / can't explain why the calculated size of the 2nd complex is much smaller the the expected size? Since I don't have the actual sequence of the Fab, could it be smaller than the general Fab size? Other than experimental error, i.e. determination of concentration, what else I have missed? It will be great if you guys can shed some light on it.

Many thanks again,
Matt

Paige
Active member
Posts: 16
Joined: 2011.10.28 04:10
Location: Hauptman-Woodward Institute, Buffalo NY

Re: SAXS MoW alternatives

#9 Post by Paige » 2012.03.02 18:57

If you divide the Porod volume by ~1.7 you get the molecular mass.
Alternatively you can take the volume from a DAMMIF (or DAMMIN) model and divide it by 2 to estimate the molecular mass.
Could you explain where these numbers come from please? Is there a reference that you can point me to? Why the difference from porod volume to the dammif excluded volume?

Thank you.

User avatar
AL
Administrator
Posts: 686
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Porod volume and DAMMIF volume

#10 Post by AL » 2012.07.19 11:22

Paige wrote:Could you explain where these numbers come from please? Is there a reference that you can point me to? Why the difference from porod volume to the dammif excluded volume?
See Petoukhov et al. 2012 J. Appl. Cryst. 45, p. 344 and supplementary materials.

molon
Member
Posts: 10
Joined: 2010.11.17 17:49
Location: AFMB, Marseille, France

Re: Porod volume and DAMMIF volume

#11 Post by molon » 2013.10.01 14:57

AL wrote:
Paige wrote:Could you explain where these numbers come from please? Is there a reference that you can point me to? Why the difference from porod volume to the dammif excluded volume?
See Petoukhov et al. 2012 J. Appl. Cryst. 45, p. 344 and supplementary materials.
Hi,

I have read the Petoukhov et al. 2012 paper and, if I understand correctly the Supplementary materials, these show roughly the same ratio for Va(dammin/dammif)/MM and for Vp(autoporod)/MM, that is, ~ 1.6-1.7. For DAMMIN/DAMMIF the ratio depends on the number of Shannon channels used, which are, in most of the cases in my hands, between 7 and 9. A lower number of Shannon channels seems to produce a higher ratio. Yet, for the examples given, at 8 Shannon channels the ratio is ~ 1.7.

Although I see in my examples that the Va/MM in DAMMIF is actually closer to 2 than to 1.6-1.7, I still don't understand why is this the case. I would appreciate if someone could indicate what I am misunderstanding in the Petoukhov et al. 2012 paper or where this discrepancy between the ratios in autoporod versus DAMMIN/DAMMIF comes from.

Thank you!

Alex
Active member
Posts: 692
Joined: 2007.08.09 21:10
Location: Planet Earth

Re: SAXS MoW

#12 Post by Alex » 2013.10.03 13:01

I have read the Petoukhov et al. 2012 paper and, if I understand correctly the Supplementary materials, these show roughly the same ratio for Va(dammin/dammif)/MM and for Vp(autoporod)/MM, that is, ~ 1.6-1.7. For DAMMIN/DAMMIF the ratio depends on the number of Shannon channels used, which are, in most of the cases in my hands, between 7 and 9. A lower number of Shannon channels seems to produce a higher ratio. Yet, for the examples given, at 8 Shannon channels the ratio is ~ 1.7.
Although I see in my examples that the Va/MM in DAMMIF is actually closer to 2 than to 1.6-1.7, I still don't understand why is this the case. I would appreciate if someone could indicate what I am misunderstanding in the Petoukhov et al. 2012 paper or where this discrepancy between the ratios in autoporod versus DAMMIN/DAMMIF comes from..
I believe that my tables would have been more complete if the information on anisometry would have been included. Anisometry (prolate or oblate object) has an effect on the numbers you get.

HTH,
Alex

molon
Member
Posts: 10
Joined: 2010.11.17 17:49
Location: AFMB, Marseille, France

Re: SAXS MoW

#13 Post by molon » 2013.10.07 11:32

Thank you, Alex.

Could you elaborate on the dependence on anisometry? Is it explained somewhere? Just trying to understand...
Thanks again!

Miguel

Alex
Active member
Posts: 692
Joined: 2007.08.09 21:10
Location: Planet Earth

Re: SAXS MoW

#14 Post by Alex » 2013.10.07 11:47

Miguel, the ratio changes with anisometry. If you have a very prolate particle, the ratio would be more like 2 (if I remember it correctly).
You can do a simple test yourself to find it out.
Alex

molon
Member
Posts: 10
Joined: 2010.11.17 17:49
Location: AFMB, Marseille, France

Re: SAXS MoW

#15 Post by molon » 2013.10.07 13:55

Thank you, Alex, I will try to do these tests. I suppose the different ratios have to do with the DAMMIN/DAMMIF algorithm. Indeed, the examples I was mentioning correspond to a prolate object.

Best,

Miguel

Post Reply