Re: [AMBER] MM/PBSA and GBSA fails

From: David A Case <>
Date: Sun, 7 Jun 2020 20:35:29 -0400

On Sun, Jun 07, 2020, Thakur, Abhishek wrote:
>I have a complex system for which I am trying to compute the binding
>affinity over 500 ns of simulation by MM/PBSA and MM/GBSA method.

>Unfortunately, the binding affinity predicted is very high, i.e., DELTA
>TOTAL = 693215.1234 Kcal/mol for one system but for other systems binding
>affinity seems to be reasonable ~ -21 Kcal/mol.

Look at the individual components of the energies, and at the individual
entries for each frame. And, of course, visualize the trajectories to
see if everything looks OK.

>An interesting thing is that ligand is the same but protein is different
>isoform and experimentally this system should be the best binder.

I feel confident that an error of 600,000 kcal/mol is not related to
what isoform of the protein you used; there is some computational error.
Fortunately, such a gigantic error is often fairly easy to find, but it
does require careful investigation of your results.


AMBER mailing list
Received on Sun Jun 07 2020 - 18:00:02 PDT
Custom Search