This question was also posted in the General Residencies Forum. I answered the question there and I've cut and pasted my response below:
Medical school reputation helps maybe a little bit but basing an application heavily on the reputation of one's medical school would be equivalent to judging a book by its cover or only looking at the tip of the iceberg. If you have a reasonable publication record and good stats, why shouldn't you have a good chance of getting into a research-oriented residency program?
I haven't seen CVs of residents being published...but I haven't really looked hard enough for them either.
Research-oriented residencies will have strong mudphud representation however there is a significant number of MD-only folks at these programs. Again, it's about the overall application rather than the letters attached to the end of your name. Now, if you have like zero publications and you're going up against all those mudphud applicants (stats being equal) who have a substantial # of publications, of course you'll be at a disadvantage when applying to research-oriented residencies.
If you're interested in research-oriented residencies, is it because you want to be a physician-scientist? If so, why not go the MD/PhD route? That's why NIH created MSTP programs...to train people like you
------------------------------
As an addendum, there is this Molecular Medicine Training program at UCSF. It is open to those who are applying to Internal Medicine, Pediatrics, Pathology, and I think Derm & Neuro as well. It basically ensures you a spot in a post-residency fellowship of your choice at UCSF at the start of residency. Also, it funds up to 3 years of your post-doc research. But here's the catch. You have to be an MD/PhD to be considered for admission to the program. Exceptions can be made for MD only applicants with strong research credentials but they clearly favor the dual-degree folks.