It's a very complicated issue and that complexity the exact reason as to why they're having such difficulty determining what causes the phenotypic expression (observable appearance of a trait) of and curing things like Alzheimer's disease.
From what little of the for-dummies stuff in Scientific American I've read lately, it appears that the real problem there is that:
A. "junk" sequences of DNA have turned out to be quite important to the formation of things that have profound effects on the final structure and function of an organism through changes in their cells.
B. It's increasingly clear that DNA isn't like computer code in the classical sense, where you put in 1 + 1 and get 2, but is more like quantum computational systems, where one system may output 2, but it'll eventually, under certain circumstances, interact with other DNA and produce 3 or 5.
Hence the specialization of our body cells is probably being caused by a few hundred different starting inputs to various genes, not just one gene doing everything that makes a blood cell a blood cell. I know that's a "duh", a blood cell isn't a muscle cell- but what's interesting is how far apart the changed DNA usually is. It's like somebody turns on a light in Brooklyn, and that causes a door to open in San Diego. But it's not butterfly effect, it's just very complex multi-input quantum mechanical machinery.
I find it fascinating that one day we'll probably have quantum processors doing more or less the same thing, and that probably soon afterwards we'll be able to do to-design organisms, so much less crude than the gene-insertion stuff that we're doing today, because we'll have computational environments that can more accurately model reality's fuzzy complexity and we'll be able to build machines that can output both the correct genes and the starting sequences needed to start the cascade.
What we're doing now, for all of it's amazing science, seems like ripping out a big chunk of a distributed computing system, sticking in a bunch of new inputs, outputs and processors and hoping that it does something vaguely useful and doesn't kill the organism.
This glow-gene thing's been applied to a lot of organisms (I remember reading about it applied to tobacco leaves at one point IIRC) so it's a fairly well-understood piece of machinery- stick here, activate thus, and it probably won't overwrite anything vital.
But I don't think we'll have glowing people soon.
Because, who knows? Maybe if you glow, you're stupid. Or you can't eat certain proteins. Or see in the full range of color. Or your skin won't make one of the chemicals it needs to be fully healthy without fur. I'm all for a future where we can actually try to improve upon Nature, but I think we're still more in the black-magic stages than anybody likes to admit.