Hi Qi,
In the latest Xplor-NIH version (2.44), the atomic radii in the default protein
parameters are those of MolProbity. This is a recent change, implemented as of
version 2.43. The new radii yield significantly fewer clashes PROVIDED THAT
they are scaled by a factor of 0.9, instead of the 0.8 of the older radii.
What's probably happening is that you are using the new radii (in version 2.43)
within a script that explicitly sets up the scale factor ("repel" variable) to
the older value, which will result in worse clashes.
There are different ways to fix this:
1) Locate the 0.8 in your script and replace it by 0.9.
2) Simply invoke protocol.initNBond() for the nonbonded setup, which
automatically recognizes the version of protein parameters being used, and
chooses the corresponding radius scale factor. This would look like this:
rampedParams.append( StaticRamp("protocol.initNBond()") )
which would set a uniform scale factor for the simulated annealing stage. You
can find an example in eginput/gb1_rdc/anneal.py.
3) Use the new nonbonded potential, which also automatically sets the
appropriate radius scale factor. You can find an example in
eginput/gb1_rdc/refine.py
Suggestions 2 and 3 are based on the latest Xplor-NIH version (2.44).
I hope this helps.
Best,
Guillermo Bermejo
________________________________
From: Hu, Qi, Ph.D. <[email protected]>
Sent: Thursday, March 2, 2017 5:46 PM
To: [email protected]
Subject: [Xplor-nih] violation in 2.43
Dear Charles and Other Users,
I used Xplor-NIH 2.39 and Xplor-NIH 2.43 for calculations. When I submitted my
structures for PSVS and MolProbity evaluations, the reported vdw violations was
much worse for Xplor-NIH 2.43 than for Xplor-NIH 2.39. I assume this is due to
the different atom radii used in the two versions of Xplor-NIH. Would you have
any suggestion on how to deal with the violations here? Should I go back to
version 2.39 for furture calculations? I have not check if the version 2.44 has
the same problem. Thank you very much for any comment.
Best regards,
Qi
_______________________________________________
Xplor-nih mailing list
[email protected]
https://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
