It is a disturbing question, in general. Such as: How much new science is being missed by brain-power sequestored in IT departments making a decent income re-writing programs to calculate taxes on teddy bear sales?

At least in theory, data scientist may be generating techniques and models relevant to multiple applications (if those applications are made or allowed to be made.) I could easily imagine population linking models being used to increase the understanding of epigenetics. My futurist projection is that the hottest area in 10-15 years will be cross-pollination.