mmpbsa/qm_mmpbsa

[maral afshinpour]

Hello, I have calculated the binding energy between two chains of protein with two methods, one with mmpbsa and one with MMPBSA(QM). In the second method, I index some residues for QM calculation,  but the binding affinity is considerably different; you can find the results attached. I am wondering whether you can help me to figure out the reason for this significant difference. Should I add some parameters to the mmpbsa? In file to calculate the binding energy while considering the QM?

Thanks,

Maral

[Mario Sergio Valdes]

Many factors affect the value of the computed energy.
First. Comparing g_mmpbsa with gmx_MMPBSA must be done with care since they are based on different calculation engines, APBS and Amber respectively. Therefore, it is likely that you will not get the same value in both. In addition, g_mmpbsa has only PB, while gmx_MMPBSA has a comprehensive combination of methods and parameters.
Considering the above, you should make sure that you use equivalent methods and parameters. According to the files you sent, you are using two different methods, PB for g_mmpbsa and GB+QM for gmx_MMPBSA. It is to be expected that there are noticeable differences. Also, you are using different solute dielectric constant, 8 for g_mmpbsa and 1 (default) for GB+QM.
QM is implemented primarily to try to reproduce hybrid simulations, however, it is optional to use it in this case (I guess). If you want to make a more coherent comparison, you should make sure, as I said before, that the methods and parameters are equivalent. Even so, you will probably not get the same results since there are innate differences, not only in methods but also in parameters (e.g. differences in radii sets).

HTH!

Mario S.

[maral afshinpour]

Thank you, based on the MMPBSA result, which number represent the binding affinity between two chain of protein?

should I consider complex part energy as a total binding affinity?

[Mario Sergio Valdes]

No. Taking into account the thermodynamic cycle it would be the Delta. See the information here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/introduction/). The value you get is very close to zero, so you may want to check cases we have discussed here in the group and on Github, where the energy is positive. One of the main and most discussed reasons for PB and GB is the total system charge and its relation to the solute dielectric constant. You should review the literature for similar systems or in an extreme case, experiment with varying the parameters logically and consistently.

Note also that this energy does not correspond to the binding free energy, since it does not include the entropic term. It is commonly referred to as effective free energy or interaction free energy.