I have a research using MD simulation (using GROMACS) for proofing the binding stability of protein-protein interaction. To measure whether the interaction is stable, I wish to employ binding affinity analysis.

I am familiar with PRODIGY, a binding affinity predictor using network of contacts at the interface of a protein–protein complex to help calculate it (https://elifesciences.org/articles/07454). However, I realized that PRODIGY is usually being used for static binding affinity and not typically used in dynamical systems.

Regardless, I have successfully employed scripts that: (1) extracted PDB files from the trajectory file (xtc), and (2) calculate each PDB file's binding affinity using PRODIGY. I plotted the binding affinity vs time and got this as a result.

Is this a correct approach to binding affinity analysis? or should I employ different analysis?

I greatly appreciate your insight and comments regarding this topic. Thankyou.