Not sure about other areas, but judging from these papers, impact of Mapreduce in biomedical field is negligible (2-3 citations for a paper is nothing). No idea why that is -- there's a clear need for parallel computation -- perhaps more traditional techniques as OpenMP see more use.
