[pattern]

To expand on your comment "rare diseases are actually quite common", I was struck by this in the article:

    More than 7,000 such diseases exist, afflicting a total of 25 million to 30 million Americans

That's almost 10% of Americans! I suppose this shouldn't have been surprising given the Birthday Paradox [1], but the ~1/10 statistic puts things into perspective for me.

[1]: http://en.wikipedia.org/wiki/Birthday_problem

[mtdewcmu]

That means most of them must be mild. Because, clearly, 10% of Americans don't have something as serious as FOP. There's probably a long tail of increasingly rare and increasingly mild diseases, so that the cutoff is arbitrary.

[waster]

Mild, or self-limiting or otherwise transitory. I know (closely) three people in two different countries who have a 15/100,000 disease that is self-limiting.

[mtdewcmu]

It wouldn't be very surprising to learn that every single person has some kind of genetic defect -- perhaps dozens -- but usually the compensatory mechanisms are strong enough to mask them, at least while you're young and strong overall. It makes a lot of sense that health should be a spectrum or a kind of dynamic equilibrium, rather than a binary on-off condition.

[jaggederest]

Yes, I'm not sure that they're diseases, per se.

I think a better term should be found, perhaps 'phenotypes' or 'variants'