How to generate the theoretical P(r) curve from pdb files?

Interactive and automated data processing tools (PRIMUS, GNOM, AUTORG).
Scattering from simple bodies (BODIES), peak analysis (PEAK), data plotting (SASPLOT) etc.
Post Reply
Message
Author
zhuwenkai
Active member
Posts: 17
Joined: 2016.09.27 19:06

How to generate the theoretical P(r) curve from pdb files?

#1 Post by zhuwenkai » 2016.12.02 09:18

Hi.generally, we get the theoretical scattering curves from high-resolution structures by CROSOL,

but i was wondering How to generate the theoretical P(r) curve from pdb files? I know a example using CPPTRAJ in the AMBER 14 package to get the P(r) curve.

By the way, one more question, i get the advice to normalize the y-axs of the P(r) curves when a series of curves are compared from the forum,

But can we normaliz all the curves to a total area of 1? that'all, thanks in advance!

cheers,
WenKai Zhu

User avatar
AL
Administrator
Posts: 675
Joined: 2007.08.03 18:55
Location: EMBL Hamburg, Germany
Contact:

Re: How to generate the theoretical P(r) curve from pdb file

#2 Post by AL » 2016.12.06 17:26

zhuwenkai wrote:Hi.generally, we get the theoretical scattering curves from high-resolution structures by CROSOL,
but i was wondering How to generate the theoretical P(r) curve from pdb files?
If you have ATSAS 2.7.2 the easiest would be to run CRYSOL and then use GNOM 5 on the predicted curve like this:

Code: Select all

crysol model1.pdb -lm 30 -ns 256
gnom5 model100.int --rmax 100.0
Here the --rmax value is 'Envelope diameter' value from the CRYSOL output.
If you want a better fit of the p(r) to your calculated curve - use smaller values of --alpha:

Code: Select all

gnom5 model100.int --rmax 100.0 --alpha 0.01
zhuwenkai wrote:By the way, one more question, i get the advice to normalize the y-axs of the P(r) curves when a series of curves are compared from the forum,
But can we normaliz all the curves to a total area of 1?
Since the area below the p(r) is equal to I(0)/4Π you may normalise all your curves by I(0) from AUTORG using DATOP DIV and then run GNOM. This way the p(r) functions should have the same area.

User avatar
anlarsen
Member
Posts: 11
Joined: 2017.03.06 16:37
Contact:

Re: How to generate the theoretical P(r) curve from pdb files?

#3 Post by anlarsen » 2018.08.19 13:20

Dear Zhuwenkai,

CaPP can calculate the theoretical p(r) directly from a PDB file (without any detour into reciprocal space):
https://github.com/Niels-Bohr-Institute ... ophys/CaPP

When comparing p(r) functions, I recommend normalising the maximum value to unity (i.e. divide by max{p(r)}), as this gives a good basis for visual comparison of the p(r) from proteins with different shapes.

User avatar
anlarsen
Member
Posts: 11
Joined: 2017.03.06 16:37
Contact:

Re: How to generate the theoretical P(r) curve from pdb files?

#4 Post by anlarsen » 2019.02.27 17:00

OBS
CaPP is now updated with several bug-fixes, and possibility for deuteration (SANS), and possibility for inclusion of explicit H and D. So please use the latest version:

https://github.com/Niels-Bohr-Institute ... ophys/CaPP

- and it's 100% open source

Post Reply