Hi Eigen, a single professor oftenly cannot decide on your admission - to have a better chance, you better fit several profs/research groups, i.e. "department". Since faculty with similar interests usually group together, that should be feasible.
By "mainstream" I mean well-established, proven, and widely-distributed research approaches, not ones that are "hot". You need to look for approaches that lots of scientists made career on - and therefore unlikely to wrap-up in a couple of years, are easily funded, and so forth..
It is much less risky to continue well-trodden path in your graduate application, especially one that fits faculty well, than try to present your own outlook. The latter you can do safely after passing the graduate admissions barrier and perhaps the first year of grad school.
Of course you can follow your heart and pick up the program closest to your research field in the 100+ range, but I bet if you really love what you're doing, you want to give it as better resources as possible, ideally in the top 10 range Even if this involves a bit of trickery.
"Ordinary" research or, how to say, research that everyone has tired of already, may actually have a better shot when it comes to graduate admissions. Of course as a scientist you need to do what you love, but this motto doesn't make much sense if you sit without lab, funding, and degree to be able to get yourself lab and funding. At this point of the career - before you're even a graduate student - the most important thing is to climb yourself up to the point where you can do what you love safely, and with enough resources to make it flourish.
----------------------------------------------
PS I don't consider this state of things in science right. But it seems to naturally follow from over-importance of "fit" in graduate admission process.
PS2 I would like to ask "haters" to stop voting all of my posts down. Thanks.