Oligomer with more than 2 pdbs and constraint file

Linear (OLIGOMER), and non-linear (MIXTURE) analysis, singular value decomposition (SVDPLOT), addition of missing fragments (BUNCH, CORAL), analysis of flexible systems (EOM/RANCH & GAJOE), flexible refinement of high-resolution models (SREFLEX)
Post Reply
Message
Author
AndreaGraz
Member
Posts: 7
Joined: 2018.03.01 17:22

Oligomer with more than 2 pdbs and constraint file

#1 Post by AndreaGraz » 2018.03.14 18:53

Hello,

I am trying to run OLIGOMER to test the ratios between more than 2 (but less than 10) pdb files, which all have the same mass and contrast,they only differ in conformation. I would like to enforce that each pdb file retains the same volume (or number) fraction.
Therefore, I run FFMAKER with the following constraint file (for example for 3 components):
Volume Fraction Relation Coefficients: 1.0 -1.0 1.0
the header of the dat file coming from FFMAKER looks like:
FormFactor file for OLIGOMER created from: toFFStructure1.dat toFFoStructure2.dat toFFStructure3.dat
Contrasts: 0.10000E+01 0.10000E+01 0.10000E+01
VolumeFractionConstrains: 1.0000 -1.0000 1.0000
0.0000000E+00 0.1253000E+08 0.1392200E+08 0.1355800E+08
0.1023600E-02 0.1252400E+08 0.1391100E+08 0.1354700E+08
However, this results in the following once I run OLIGOMER (with constant)
Using constant as additional component
experimental.dat 62.86 3311 41.50 0.000+-0.000 0.488+-0.002 0.488+-0.002 0.024+-0.000
This happens on both ATSAS-2.7.2 and ATSAS-2.8.3

Two of the 3 pdbs do indeed have the same population, but the third one is set to zero. How do I enforce A=B=C...=N? I have tried setting the volume fractions to, say 0.33 each, but this makes no difference. Is there a way to check the scattering for N components when each component makes up 1/N of the sample?

Many thanks!

konarev
Administrator
Posts: 52
Joined: 2007.09.14 15:38

Re: Oligomer with more than 2 pdbs and constraint file

#2 Post by konarev » 2018.03.16 14:31

If the task is only to calculate the fit from N component with equal volume fractions,
then one does not need to run OLIGOMER with the form factor containing multiple components
as there is no need in optimization of their volume fractions, the volume fractions are already known and fixed.

In this case, one has to calculate the individual theoretical scattering patterns from
pdb models (e.g. using CRYSOL), average them (e.g. using DATAVER or PRIMUS) and use this curve
as a single component in the input form factor file for OLIGOMER.
OLIGOMER will properly scale the theoretical averaged curve to the
experimental data and provide chi^2 value (fit quality).

AndreaGraz
Member
Posts: 7
Joined: 2018.03.01 17:22

Re: Oligomer with more than 2 pdbs and constraint file

#3 Post by AndreaGraz » 2018.03.16 16:58

Thank you very much for the advice!
I will average the theoretical curves with DATAVER and use the output for oligomer then...

AndreaGraz
Member
Posts: 7
Joined: 2018.03.01 17:22

Re: Oligomer with more than 2 pdbs and constraint file

#4 Post by AndreaGraz » 2018.07.27 12:41

Thank you once again for the help on this topic. Running DATAVER and then feeding oligomer a single form factor does indeed work for volume fractions, but I am now tasked to do a similar thing but with molar fractions.

Running DATAVER on the raw cryson/crysol int files does not work in this case, as the I(0) of these files is scaled to the volume in solution. I was wondering if there is a number in the cryson/crysol log file that I can use to normalize the int files so that the I(0) is now appropriate for the molar fraction?
As these are just pdb files of multiple conformers, not different oligomeric states, can I just scale the int files to all have the same I(0) and then average? Would the result then be a molar fraction?

Sorry for the strange question... I have an equilibrium of multiple conformers by NMR (with NMR data depending on molar fraction) and I am trying to see if the scattering data could make sense with it....

Many thanks.

konarev
Administrator
Posts: 52
Joined: 2007.09.14 15:38

Re: Oligomer with more than 2 pdbs and constraint file

#5 Post by konarev » 2018.08.01 16:42

I don't think that Crysol/Cryson have an option to normalize the scattering curves.
For different NMR conformers I(0) should be the same, so you can scale them if they have different I(0) and then average.
Also you can try to use an option '-cst' for oligomer (if you have not used it yet), in this case a constant will be used as an additional component,
in some cases it can improve the fit quality at higher angles.

AndreaGraz
Member
Posts: 7
Joined: 2018.03.01 17:22

Re: Oligomer with more than 2 pdbs and constraint file

#6 Post by AndreaGraz » 2018.08.01 19:03

Thanks for the advice. Indeed, as some are more extended than others, the I0 of the int files from crysol differs quite a bit (up to 40%), presumably because of the difference in solvent shell.

So I will then scale them all to each other using the I0 from AUTORG and run DATAVER as suggested above, in order to get the theoretical scattering of conformers with equal number fraction.

Post Reply