QM Region

[Nathaniel STILLSON]

Hi,

I am attempting to perform QM/MMGBSA for single trajectory protein-ligand simulations. While trying to establish exactly what region to treat with QM I came across some papers that only treated their the ligand with QM , not any residues from the receptor (e.g this paper). I attempted to do this with my system (using igb=5, qm_theory=PM3) and got a non-sense result (∆G=-20000 kj/mol) along with the warning:

GENERALIZED BORN:

WARNING: INCONSISTENCIES EXIST WITHIN INTERNAL POTENTIAL TERMS AND

THE VALIDITY OF THESE RESULTS ARE HIGHLY QUESTIONABLE!

Some absolute differences in the internal potential terms are greater than 0.005.

This should not happen when using Single Trajectory Protocol!

This error goes away once I include a residue from the receptor in the QM region. Do I have to include a residue from the receptor in the QM region for a reasonable result, or could there be another source for my error?

Thanks kindly

[Mario Sergio Valdes]

If you don't include the QM calculations, do you still get the warning? Do the results make sense to you when you include the residues from the receptor? 

[Nathaniel STILLSON]

I do not get the GB error when doing the calculation without QM. Likewise, when residues from the receptor are included in the QM region, i do not get the GB error. I have looked at two different QM regions with residues from the receptor: one with a single residue that makes an H-bond with the ligand, and another with all residues "within 5.0" these two differ significantly (~ -24 kcal/mol and ~ +30 kcal/mol) and neither are particularly close to GBSA result with no QM (~ -53 kcal/mol). 

Thanks again

[Mario Sergio Valdes]

Can you show me the output files (FINAL_RESULTS_MMPBSA.dat) for each of the cases?
According to the reviews concerning the subject, QM has not proven to improve considerably the calculations in all cases, so its use should be with consideration.

[Nathaniel STILLSON]

Ive attachted 3 output files. One example of a regular GBSA calc, one where only the ligand is added to the QM region (residue 286) and one where all residues within 5.0 are added to the QM region.

Thanks

[Mario Ernesto]

It seems that it's necessary to include the ligand and one residue for the receptor at least, otherwise, everything falls apart... when the 'ifqnt' option is on, sander uses a different algorithm designed for Hybrid-Potential calculations... by including the ligand and at least one residue from the receptor, it is ensured that the same algorithm is used for the three components (i.e. complex, receptor, and ligand) and everything matches up...

hope this helps!

--
You received this message because you are subscribed to the Google Groups "gmx_MMPBSA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gmx_mmpbsa+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gmx_mmpbsa/2a9ab3c8-5b3a-40d9-a057-db7cdb399997n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Nathaniel STILLSON]

Thanks for clearing this up for me! I'll have to perform the calculations with some residues from the receptor in the qm region them then