Is there any advantage in considering more images? Where?

Ab initio modelling: DAMMIF, DAMMIN, GASBOR, MONSA
Post Reply
Message
Author
Albert
Member
Posts: 5
Joined: 2017.11.29 23:48

Is there any advantage in considering more images? Where?

#1 Post by Albert » 2017.11.30 01:07

Dear SAXS users!

Can somebody help me? Consider two independent datasets from the same protein (at the same concentration) consisting of the average of 10 images for one and 300 images for the other. Both datasets had Qmax=0.4 Angs^-1 (Rg=16 nm) and were free from radiation damage. Find attached an image.



My question: Is there any advantage in considering more images? Where?

I have tried to see if considering more images leads to better unique models by checking the NSD values (with Damaver), the resolution (with Sasres) or the C/p-values in the CorMap test. I have not been able to conclude something in this direction.

Here are some of my results:

DAMMIF(20 independent models):
    with 10 frames --> NSD=0.556+-0.013, Res=18+-2
      with 300 frames --> NSD=0.495+-0.011, Res=17+-2

      DAMMIN(10 independent models):
        with 10 frames --> NSD=0.505+-0.005, Res=20+-2, C/p-value=21/0.000277
          with 300 frames --> NSD=0.523+-0.007, Res=20+-2, C/p-value=16/0.008828

          Bests regards,
          A.
          Attachments
          image_1.png
          image_1.png (10.16 KiB) Viewed 5618 times

          Post Reply