Protocol for obtaining the ab initio model

Ab initio modelling: DAMMIF, DAMMIN, GASBOR, MONSA
Post Reply
Message
Author
nanolife
Member
Posts: 5
Joined: 2023.11.27 22:43

Protocol for obtaining the ab initio model

#1 Post by nanolife » 2023.11.30 16:07

Dear SAXier,
I am having problems determining which would be the best representative model for my SAXS data.
Here is the flow of my work. The SAS data processing performed Dammif to get 12 bead models. The bead models were then clustered into different groups with the most representative model selected. It also performed a global comparison of all the models, picked out the most representative global model, and then from here it automatically refined with Dammin to get the final model. However, my final model with Dammin has fewer dummy atoms (~ 200 atoms while the Dammif model has ~ 1500 atoms) and cannot be presented in mesh mode. I looked at the log file of Dammin, it said that the model was run with ~5000 starting beads.
Then I tried to run Dammin alone and had the same problem. I think there is a problem with the new version of Dammin which did not happen with the 2.8 version. I now can't access the 2.8 version.
Coming back to my problem, since I could not get my final refinement data, I had a difficult time selecting which most representative cluster as my model because the representative models from each cluster are so different. This is a new molecule so I don't know which model I should use so that I don't have any bias.
My question is what could be the best way to select an ab initio model for SAXS data?
Thanks,

User avatar
Guðmundsdóttir
Active member
Posts: 17
Joined: 2008.10.13 17:54
Location: Reykjavík

Re: Protocol for obtaining the ab initio model

#2 Post by Guðmundsdóttir » 2023.12.05 10:30

nanolife wrote: 2023.11.30 16:07Coming back to my problem, since I could not get my final refinement data, I had a difficult time selecting which most representative cluster as my model because the representative models from each cluster are so different.
If the Dammif or Dammin models look different but fit the SAXS data equally well then there is no way to say which one is "best". For details see e.g. Petoukhov & Svergun (2015) Ambiguity assessment of small-angle scattering curves from monodisperse systems

User avatar
AL
Administrator
Posts: 821
Joined: 2007.08.03 18:55
Location: Hamburg, Germany
Contact:

Uniqueness of ab initio shape determination in SAS

#3 Post by AL » 2023.12.06 20:24

See also Volkov VV and Svergun DI Uniqueness of ab initio shape determination in small-angle scattering J. Appl. Cryst. (2003). 36, 860-864

Post Reply