ChromosomesPhoto: © Howard Sochurek / CORBIS

"We've always known that the X chromosome was important for disease, and that there are a great many X-chromosome-linked diseases in males," he says. Such rare genetic diseases more likely strike men, because a disease-causing mutation on the lone male X chromosome cannot be compensated for by a protective normal gene on the paired X chromosomes of women. "What fascinates me about these new studies is that they may give us an insight into far more common disorders that show characteristic differences in the frequency between males and females.

"Autism, for example, is about four times more common in males than in females. Why? Rheumatoid arthritis and many other autoimmune disorders are much more common in females than in males. Why? Our results at least raise the possibility that these genes are failing to be fully dosage-compensated, creating a characteristic dosage difference between males and females. And those genes could likely play a role in increasing or lowering the susceptibility of one sex compared with the other to some of these conditions.

"But we also have no idea whether the variation is the same in a fetus in utero or in a newborn, or in a ninety-year-old woman," says Willard. "And it may be that gene expression is changing during that time, and that change may associate with late-onset diseases such as heart disease."

The new findings are only the latest emerging from decades of Willard's research on the phenomenon of dosage compensation. His scientific quest began as a Eureka! moment he had as a Harvard undergraduate. "I was sitting in a library flipping through a journal waiting for a professor who was late. And I came across this paper on X inactivation, and it just struck my fancy. The basic lesson from this paper was, 'we haven't a clue what's going on here.' To me, that was the greatest way to enter a scientific problem, because your imagination can run wild. People were just shrugging their shoulders and saying, 'It makes intuitive sense why males and females would need to equalize dosage of genes.' It was as if a 'miracle' occurs, and it just happens."

Willard recalls that he became instantly fascinated by the scientific mystery of this unknown biological mechanism, which is central to the development of every female. "I must have written every one of my papers as a biology undergraduate on this topic. And I kept reading and writing and exploring different models in my mind." The uncharted machinery of dosage compensation resonated with his (perhaps genetic?) predilection for black-box problems. "I've always been bored by projects where the answer was too obvious--where the answer was going to be one or another known possibilities," he says. "I found it much more interesting to dream about possibilities that just hadn't been described yet."

The fascination endured. When Willard started his own research laboratory after receiving a Ph.D. in human genetics from Yale University in 1979, his first goal was to figure out the machinery the cell uses to shut down X-linked genes during embryonic development. Fifteen years of painstaking work led to the identification of a master genetic switch that turns off such X chromosome genes. But this switch was a peculiar gene, indeed. The huge majority of known genes are blueprints for "messenger RNA" that produces proteins; however, this gene, dubbed XIST, instead produces a type of RNA that controls other genes. These genes that control other genes represent the next great frontier in genomic research, Willard says. Traditionally, geneticists have focused on how genes code for proteins; now they are beginning to explore the "epigenetic" machinery by which genes themselves are controlled.

As he delved into the machinery of X inactivation, he encountered other surprises. The X chromosome control system did not function as a single on-off switch, like the master circuit breaker in a house. Rather, Willard was to discover, it acted more like the multitude of individual electrical switches within that house, with different switches for different genes. He recalls the first inkling he had that X inactivation wasn't an all-or-nothing proposition. "I was teaching an undergraduate class at the University of Toronto back in the late 1980s, and I assigned students what I thought was a simple little project--to look at gene expression on the X chromosome. And the students came up with an answer that made absolutely no sense at that time. They found a gene still being expressed, even though it was on the inactive copy of the chromosome instead of the active copy. And even though I was tempted to simply say, 'You're wrong. It can't happen,' and put a big X across the lab report, we started looking into that question."

At that point, Willard's black box transformed into a treasure chest. He and his colleagues discovered a dozen examples of genes that escaped silencing. In recent years, as the Human Genome Project has yielded the complete structure of the X chromosome, the researchers have used that knowledge to find hundreds more.

They are now exploring not only how the cell decides which genes should escape silencing, but also, why. And they are seeking the origins of the startling variations they discovered among women in the genes that escape silencing. "Maybe the patterns are random, but it's much more intriguing to me to consider that the pattern of this gene activation is inherited," Willard says. "If so, when we compare the X chromosomes of mothers and daughters, or of sisters, or of identical twins, we should see a familial pattern. If it is a pattern in the genome, then we're off on another hunting expedition. Somewhere amidst the vast stretches of DNA on the X chromosome there is some sequence of DNA that tells those genes to be expressed or not expressed. It's another genetic code that we don't understand and can't even begin to articulate."

If the machinery of X inactivation is a fascinating set of nested black boxes, Willard's other major research object, the centromere, has proven a murky Stygian nightmare. The centromere--the point at which paired chromosomes are attached to enable them to navigate through cell divisions--had been largely shunned by scientists, because it was thought to be a genomic wasteland. It seemed to be nothing more than genetic "stuttering"--regions of inanely repetitive DNA code that had no purpose other than to take up space and frustrate biologists.

In fact, the so-called "complete" sequencing of the human genome, announced with great fanfare in 2000, did not include any sequences of the seemingly unfathomable centromeric regions. Willard recalls that, in the 1980s, "there was this series of wonderful papers arguing over whether all this repetitive DNA should be called 'junk,' 'garbage,' or other pejorative terms. But I just sort of took it on faith that nature wouldn't do that. This is 5 percent of the entire genome--a stunning amount to be unimportant and just sitting in a garbage heap at the center of the chromosome."