In the 12 years since the Human Genome Project was completed, biologists
have linked more than a thousand regions of the genome to disease. “But
in most cases, we don’t actually know how they function,” says Manolis
Kellis, a computational biologist at Massachusetts Institute of Technology.
Enter the epigenome. If the human genome is the book of life, the epigenome
is the collection of bookmarks and highlighting that tells the cell what passages
of the book to read. These marks include chemical tags on DNA that make genes
unreadable, as well as chemical tags on proteins that help expose DNA inside
the cell nuclei, making genes readable. They’re the reason that cells from the
liver, heart or brain differ profoundly. The National Institutes of Health
Roadmap Epigenomics Consortium, including Kellis, published the
most comprehensive map of the human epigenome in February.
Kellis led the data analysis team, which applied machine-learning
algorithms to decode the language of the epigenome. The map serves
up important clues about how a single fertilized egg can develop into
the diversity of tissues in the human body — and how healthy tissue can
become diseased. For example, one team in the consortium reported how
metastatic cancer cells contain an epigenetic fingerprint that reveals the tissue
they came from, which could lead to better-targeted cancer treatments. Another
team reported spotting DNA sequences that may trigger autoimmune diseases.