Researchers believe something called the Epigenome Roadmap a cascade of two dozen papers published last year is gradually changing that. The Roadmap is a catalog of the millions of epigenetic switches that control gene action. It's pretty clear that epigenetics is key to really understanding disease as well as normal human traits. Eventually, that is; nailing down those connections is going to take time.
For one thing, disease mutations and other DNA variations, it turns out, are hardly ever where you would expect them to be: in genes that make the proteins that run the show. Ninety percent of disease-related mutations are in the regions of DNA that lie outside those genes, the regulatory regions that control protein-coding gene action. The working hypothesis is that variation in disease susceptibility - or any other trait - depends mostly on subtle differences in the expression of protein-coding genes, which is under epigenetic control.
Mutations associated with Alzheimer's disease, in a surprising and immediately practical finding among those new papers, turned out to be active not so much in brain cells, where you might expect activity. Instead, they altered epigenomic activity in cells of the immune system.
Comment: Video on Hawking's remarks: