Hi, I am designing a protein interface of a nanobody with my target antigen.
1000 initial poses were generated by patchdock and 1000 designs were generated per each pose.
After ploting the ddg-rmsd scatter plot, I found all the poses has a characteristic scatter plot like fig. A attached (a narrow RMSD range within ~0.25A, while a majority stays exactly at 0.00A).
The design method is generally: dock -> design interface residue within 10A (x 3 times, with minimize and repack after each round) -> final minimize and repack (x 100 times) -> calculate ddg and rmsd
Since I have also got several "more funnel-like" scatter plots by using the same design method on known nanobody-antigen complex structure (fig. B), I wonder if it is really the problem of my design script or/and the problem of the pose.
I am guessing it is too early to discard the pose from fig. A because somehow it looks like the tip of funnels I have seen in other research of interface design.
Then I want to know, is there a way of sampling a wider range of RMSD of poses by Rosetta Dock? Do you think it helps?
Thank you so much!!!
| Attachment | Size |
|---|---|
 fig_AB.jpg | 138.91 KB |
Solved! I just used the low-resolution (centroid) mode of Rosetta Dock and it is now sampling a wider range of RMSD!
Remarks:
Originally I thought full atom mode was fine but now I am hypothesizing for poses with high complementarity score from patchdock, the strong stabilization at RMSD=0.00A (i.e. exactly the same pose) by the excessive refinement of the full atom mode may have prevented the sampling of more diverse poses (i.e. wider range of RMSD) around the initial pose. However, sampling around the initial pose is important for the assessment of the folding funnel. If you meet a similar situation, just switch to the low-resolution (centroid) mode of Rosetta Dock.