Bioinformatician here: First of all, I think it's got to be clear that a lone bioinformatician, or even a group of them isn't going to go about changing the world. Essentially, when you sign up for this, you're still a cog in a machine, albeit a slightly more altruistic machine.

Here's my pet peeve in bioinformatics - If there's one thing that's poorly suited to science, it's the building of computational infrastructure. We're talking basic stuff like databases, tools etc. Sure, anyone can knock out a bit of code for a basic database, but the big problem is that there's no incentive to have decent code, or maintain it so that it lasts any longer than the person is in the lab, or has funding. So, what will be great is if existing resources are cleaned up - data is normalised and pulled out so that it is actually accessible for doing some kind of analysis on it.

If you want to do bigger work, do something actually novel, or that has any biological relevance there's no getting around collecting your own data (e.g. sequencing the crap out of a bunch of things). I'm in the process of trying to get funding now for a project of mine to make that very leap now.

I'm sure someone working on next gen sequencing (the new hotness) can pipe up with the big problems to be solved there.