CRYSOL - strange behaviour with long molecules

[polysaccharideGuy]

Hi everyone, I've been using CRYSOL to calculate the SAXS for linear polysaccharides. I'm seeing some odd behavior as the length of the molecule is increased. I've tried to test this by calculating the scattering while increasing the length of my molecule, from 10-100 monomers in 10 monomer steps. Below is the result when using CRYSOL 2:

Crysol2DPSeriesB.png (55.95 KiB) Viewed 4850 times

Once normalized for the number of monomers in the system, the high q scattering decreases with increasing molecule length above a length of ~12 nm (30 monomers). As I'm simply making my molecule longer - and not changing the conformation - I would just expect the intensity to shift up by some multiplicative factor. My first thought was that perhaps there's some (molecular) size limit in the algorithm which I'm exceeding, which based on where I start to see the issue would be about 12 nm (approx 4 nm per 10 monomers). Things get even more confusing when repeating the same steps using Crysol 3:

DPSeriesCrysol3B.png (157.47 KiB) Viewed 4850 times

Again, for 10, 20 and 30 monomers the scattering is what I'd expect (3 bluest curves). However above 30 monomers, the scattering changes drastically.

If anyone has some advice that would be greatly appreciated. I've looked through previous posts and haven't seen anything relevant, although I realize my issue might be something basic I've overlooked. Thanks in advance, Ben.

[garicman]

Hello,
Have you found a solution to your problem? I have a similar problem. I feel like CRYSOL does something with boundaries of the space occupied by the molecule

[d-sv]

Hello, this might be due to the number of harmonics used. You may attach your models to check.

[mawhite]

Hello,

This may be out of date, but the original Crysol documentation found that the solvent approximation used by Crysol breaks down above about s~0.5 (Å^-1). Your curves show data beyond s>1.5 (Å^-1) where more complicated solvent models are required. However, the effect in your curves is seen below s<0.5 (Å^-1), so I would guess that the number of harmonics is also important. Although, with very long molecules, with very high aspect ratio's, such as polysaccharides and amyloids, the harmonic approximations will probably always be insufficient.

Code: Select all

crysol Amyloid_Fibres_*.pdb SAXS.dat -lm 50 -sm 0.5 -ns 151;

I found that even using -lm 50 has similar problems as the amyloid fibres increase in length. I may be wrong, but I never use Crysol beyond s=0.5 (Å^-1).

Best regards,
Mark