Problem with molecular path is that everything seems to be getting more and more esoteric (i.e. the purview of the reference lab or large academic center). As you get more you may get more useful for individual patients but you also get less useful for the general population. You need one machine plus multiple techs, etc, to run a factor V leiden assay which is relatively "simple" molecularly (and that's just one test). An important test but no way anyone who would get a couple dozen a year could ever compete with a big center that gets a couple dozen a day (because of costs and the simple fact that the more of those tests you do the better you are).
I think (I may be wrong) molecular might be a nice fellowship for someone who wants to be in academics but doesn't necessarily want to do molecular as a career - just to do it, be familiar with it, use it in research, etc. But then again, you probably don't need to do a fellowship to get what you need.
It seems logical that as molecular tests become more prevalent and more useful and important for diagnosis that private labs will start doing them and need fellowship trained people to run them, but probably not to a huge extent.
