I think there are two main options: you can either design your experience using only the “common subset,” or you can design interfaces for multiple different input mechanisms and pick the best one based on what your user is using at that time. For the first option, pretty much every platform has a concept of clicking (single pointer), so if you design your entire interface around the idea of one-pointer clicks, that should work for every platform. Pretty constraining, but very straightforward.
If you want to be able to use platform-specific interactions (mouse-wheel, gesture zoom, hover, long-click, etc.), they you’ll probably need to implement multiple input interfaces and swap between them based on what the user is doing. This is much more laborious, but it may yield a much better user experience, so you’ll have to decide if that tradeoff is right for your scenario. Regarding how to implement these multiple input mechanisms, I know @PolygonalSun has been working on some cool stuff on this axis, but I’m not 100% sure of the details or current state.