How to Be a Computational Biologist?

In the machine learning world, a common way to operate is down load some one else's data set and try to beat the state of the art by x %. But in biology, when the goal is to discover something, no right to explain something, right? So if you've got the itch to be involved with discovery, not methodological development, but with discovering what's true about life on planet earth, how does somebody with the computational bent dive to this world sut's a really interesting thing.