That would be a difference of 28 kcal/mol, which is a lot. Is it really the full binding free energy or just the Coulombic term? What type of mutation?

See section 3.3.4 and Figure 3 in:

https://pubmed.ncbi.nlm.nih.gov/29091991/

Thank you, Martin Klvana. Yes, this is final, binding free energy. And this is a single mutation; LYS is mutated by GLU.

I used trajectory (after doing MD simulation ten ns ) in MM/PBSA.

Arunima Verma

The electrostatics plays the major role, and this might be problematic with MM/PBSA. Try to confirm the MM/PBSA value with alternative binding free energy methods, such as linear interaction energy or free energy perturbation, which include the solvent effects explicitly.

Also, run several parallel MD simulations and average the results. A spread lower than 2 kcal/mol is an indicator of good convergence.

Thank you, Martin klvana

Could you please share any good paper and tutorial regarding the free energy perturbation method for protein and DNA systems?

Although I did gromacs tutorial (FEP).

And I did not understand PARALLEL MD SIMULATION.

What dielectric constant are you using?

I used dielectric constant 8 Abdul Basit

You can check different dielectric constants. The free energy value also depends on the length of simulations and the number of selected frames from a region. Please check the latest MM-PBSA papers for similar systems.

Thank you, Abdul Basit. yes, I checked the paper and I found is 8 is relevant for protein and DNA systems. I tried different dielectric constants as well.