In my opinion, if you just want to check complex interfaces, you may do that in a qualitative manner, maybe using an empirical method. To make an accurate quantitation there are too many variables and you will probably need a polarizable force-field, even go to QM. You need to account for ring currents, Van der Waals changes, and many other microscopic interactions.
The analysis of a trajectory in this way may be a very hard job for the computer.
You can calculate chemical shifts based on your MD trajectories using some softwares, such as shiftx2, SPARTA+, and then calculate the CSPs. Compare these CSP results with experimental CSP data.