Any known example of research project being led by statistician in public health (I am fully aware that my topic may not be relevant here, do let me know if this is true.)
I am a research staff with training in epidemiology and biostatistics, working in government organization on public health matters like disease surveillance and infection control. Being a research staff here keeps giving me a feeling that I cannot lead a research project in this position, because I lack the context needed to ask sensible research questions. Doctors/PhDs in our service, although they know little about statistics, can become principal investigators of research projects, and then assign research staff to handle the technical parts.
I wonder if this kind of "statistician can only take supportive role" curse is universal to other fields also? Any comment or counter argument on this, and any suggestion of breaking this kind of curse?
 A: I don't see it as a curse.  I think in most cases statisticians are not the subject matter experts, hence they lack the deep knowledge to identify the important problems and the expertise needed to tackle them.  Whereas, the subject matter experts usually lack the deep statistical knowledge required to solve those problems.  Becoming aware of those strengths and weaknesses necessarily lead to collaboration. The best collaborative efforts are usually those where the investigator seeks the statistician's input early on. 
Great collaboration usually lead the statistician to explore and dig deeper on aspects of the project beyond what the original question was. In fact, that's how many new methods are initially thought of. Cox's proportional hazard is probably the best example illustrating how collaboration leads to new ideas.
There are also rare examples of professionals crossing bridges. Some epidemiologists are excellent statisticians (e.g. Sander Greenland) and I've also seen some statisticians becoming epidemiologists, as well as social scientists beoming excellent statisticians (Rand Wilcox, Paul Allison).
As someone famously said, statisticians get to play in everyone's sandbox.  A corollary is that in many cases, we do not own the sandbox.
