When Bradley Bernstein came to Boston from Seattle in 1999, he had planned to spend just a few years on the East Coast to complete his medical residency training at Brigham and Women's Hospital and do a postdoc in the lab of HHMI investigator Stuart Schreiber at Harvard University.

Schreiber's group was developing cutting-edge genomic technologies, and Bernstein soon found himself rubbing elbows with some of the top scientists working in systems biology, a newly emerging field that studies organisms or biological phenomena as whole systems rather than breaking them down into individual genes or proteins. Boston, it seemed, had more in store for him than he had anticipated.

Bernstein saw an opportunity to use systems biology to extend the study of protein structure he had pursued in graduate school to understand how DNA is packaged and organized inside cells. In the nucleus, DNA winds around a protein scaffold that keeps the DNA organized and regulates whether genes can be read and expressed. The resulting complex, called chromatin, can take several forms, but just how many remains a key scientific question. For the most part, however, chromatin is found in either a repressed state, which prevents a given DNA sequence from being read and therefore expressed, or an active state in which the DNA is freely accessible and able to be read.

Bernstein envisioned that being able to scan the chromatin state of a particular DNA sequence would give scientists insight into how the structure of the genome influences which genes are expressed in a cell at any given moment.

At the time, most scientists studied chromatin on the scale of a single gene or a handful of genes at most. But Bernstein thought that it might be possible to use a genome-wide approach to look at the big picture—to see whether a single snapshot of the whole cell's chromatin state would reveal the genetic program that helps determine the developmental fate of a given cell. As a postdoctoral fellow, he took major steps in achieving his long-term goal when he applied microarrays to test the chromatin state for thousands of genes at a time, first in the model organism yeast and then in human cells.

"We can now lay out the whole-genome DNA sequence and then, on top of that, add an additional map that depicts how those sequences are organized into chromatin," says Bernstein, who is now at Massachusetts General Hospital. He calls this additional layer of information the "epigenome," and it is revealing how inherited changes in gene activity can result from modifications in the way genes are structured as opposed to alterations in the primary DNA sequence.

During growth and development, genes that should not be expressed are physically tagged with chemicals, such as methyl groups. Genes can also be silenced by modification of the chromatin or histone proteins that make up the "smart stuffing" of chromosomes. These chemical modifications can influence the expression of genes even though they are not part of the actual gene sequence. In certain cases, the chemical modifications can be inherited along with the gene to play an important role in shaping the development and fate of the different cell types in the body. This kind of inheritance is termed "epigenetic," as it does not involve changes in the genetic sequence.

"These epigenomic maps show us how various cell types use their genetic information in such different ways," says Bernstein. For instance, he can compare the genetic program active in an embryonic stem (ES) cell, which retains the ability to become any cell type, with a more differentiated cell such as a neuron.

Recently, Bernstein and his colleagues discovered a key epigenetic feature of ES cells that seemed counterintuitive at first. In fact, the results initially seemed so far-fetched that Bernstein thought he and his technician had made an error. They were running an experiment to identify which genes in ES cells had repressive chromatin elements and which had activating elements. When the technician showed him results indicating that several genes had both, Bernstein told her they must have done something wrong. When a repeat experiment showed the same result, Bernstein again chalked it up to a misstep.

But in a microarray experiment testing the same question across the whole genome, "we saw hundreds of instances of the same thing—repressive and activating elements present on the same gene sequence," recalls Bernstein. "And these weren't just any old genes." They were master regulator genes, the gatekeepers of different developmental pathways. Bernstein called these opposing elements "bivalent domains," and he theorized that they keep the master regulator genes shut off in ES cells but poised to turn on quickly when a cell needs to differentiate.

The experience made Bernstein a firm believer in the power of new ideas and technologies that help scientists ask questions about what's happening inside cells on a whole genome scale. "When you can create a map across chromosomes and have thousands of data points, these trends jump out at you, and it's a very powerful approach to science."

Bernstein has built a multidisciplinary research team in Boston that will help him push these technologies to the next level—mapping the whole human epigenome. "I have benefited from surrounding myself with collaborative, very capable people," says Bernstein.