MD GIFs: keep view coordinates when ligand has drifted

[JenkeScheen]

This kind of stuff happens quite often we run MD on weak binders. I suspect that during dataviz there's a step that focusses the view on the ligand, instead we should just be sticking to the canonical pocket view.

[hmacdope]

I think what is happening is that the view is maintained but the simulation is drifting, we need to align the simulation to the canonical coordinates, (should be easy by RMSD imaging, e.g in MDAnalysis.)