Questions about appropiate usage of gmx_MMPBSA

[Gustavo Avelar Molina]

Dear gmx_MMPBSA users,

I'd like to use MM/PB(GB)SA to calculate and compare the binding energies of one ligand with several (structurally similar) wild-type proteins and mutants of one of these proteins. Coincidentally, most of the mutations I have studied experimentally are alanine mutations, which makes gmx_MMPBSA an even more interesting candidate for this study. However, before I spend the time using the tool, I'd like to understand some things:

1. In the requirements page, it says that I need a "Final GROMACS MD trajectory, fitted and with no pbc". Here, does "no pbc" mean a trajectory whose pbc effects have been removed? And by "fitted", does it mean that I should process the pbc-corrected trajectory with "-fit rot+trans"?
2. In the paper, it says that Single Trajectory Protocol (STP) "...is valid when the bound and unbound states of the receptor and the ligand are similar". But what does "similar" mean here? Structurally similar? Does it mean that, if I have a long ligand (like the grey molecule in the figure below), this method may not be appropriate because the ligand binds at a conformation that is far from the average in the unbound state? In that case, would it be preferable to use Multiple Trajectory Protocol (MTP)?
3. Is there any recommended simulation length and/or number of frames to extract in total or per nanosecond with gmx_MMPBSA?
4. For the usual MM/PB(GB)SA calculation, is gmx_MMPBSA generally suitable for good working laptops (e.g., I have dedicated 4 cores and 8 GB RAM to an Ubuntu virtual machine), or is an HPC cluster recommended?

I appreciate any help you may provide.

Best regards,

Gustavo

[gmx_MMPBSA]

Q: In the requirements page, it says that I need a "Final GROMACS MD trajectory, fitted and with no pbc". Here, does "no pbc" mean a trajectory whose pbc effects have been removed? And by "fitted", does it mean that I should process the pbc-corrected trajectory with "-fit rot+trans"?

A: Yes, more info here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/gmx_MMPBSA_running) 

Q: In the paper, it says that Single Trajectory Protocol (STP) "...is valid when the bound and unbound states of the receptor and the ligand are similar". But what does "similar" mean here? Structurally similar? Does it mean that, if I have a long ligand (like the grey molecule in the figure below), this method may not be appropriate because the ligand binds at a conformation that is far from the average in the unbound state? In that case, would it be preferable to use Multiple Trajectory Protocol (MTP)?

A: Since you are interested in relative binding free energy differences, STP is your best choice... Less computation, and error. There is plenty of literature out there about STP vs MTP. Some to get started here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/introduction)

Q: Is there any recommended simulation length and/or number of frames to extract in total or per nanosecond with gmx_MMPBSA?

A: Usually, a region where the energy is stable is preferred. This paper (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0137787) describes well the procedure to choose the proper interval. 

Q: For the usual MM/PB(GB)SA calculation, is gmx_MMPBSA generally suitable for good working laptops (e.g., I have dedicated 4 cores and 8 GB RAM to an Ubuntu virtual machine), or is an HPC cluster recommended?

A: Depends on the system size, number of frames, and method used to calculate the binding free energy. Laptop should be fine using STP, and GB, combined with "mutant_only=1" in the alanine scanning input file to avoid the calculation of the native system multiple times. More info here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/input_file/#alanine_scanning-namelist-variables) and here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/examples/Alanine_scanning)

All and all, if you are interested in performing alanine scanning or binding free energy calculations with implicit solvents, our software provides more robustness and flexibility. For a thorough comparison with other software used for similar purposes, check here (https://valdes-tresanco-ms.github.io/gmx_MMPBSA/dev/versus)

Hope this helps!

Let us know if you have further questions...

gmx_MMPBSA team

[Gustavo Avelar Molina]

Dear gmx_MMPBSA team,

Thanks a lot for your helpful reply. Now I have gmx_MMPBSA running nicely. :)

I have another question, but because it is more specific, I will make a separate post for that.

Best regards,

Gustavo