Junk DNA surprisingly active

The new work also overturns the conventional notion that genes are discrete packets of information arranged like beads on a thread of DNA. Instead, many genes overlap one another and share stretches of molecular code. As with phone lines that carry many voices at once, that arrangement has prompted the evolution of complex switching, splicing and silencing mechanisms -- mostly located between genes -- to sort out the interwoven messages.

Perhaps most surprising was how much of the human genome is at work at any given time, the scientists said.

Researchers have long known that only about 2 percent of human DNA is involved in making proteins, the molecular workhorses inside cells. That involves a two-step process in which a stretch of DNA -- a gene -- serves as a template to produce a strand of RNA, which is then used as a template to produce a protein.

Recent studies had shown that some snippets of DNA between genes also are transcribed into RNA even though they do not go on to make proteins. Surprisingly, though, the new work shows that most of a cell's DNA gets transcribed, raising questions about what all that RNA is doing.

Some of it may be doing nothing. "It may be like clutter in the attic," Collins said, noting that clutter could be useful when conditions change and evolution needs new material to work with.

But much of it seems to be playing crucial roles: regulating genes, keeping chromosomes properly packaged or helping to control the spectacularly complicated process of cell division, which is key to life and also is at the root of cancer.

Another aspect of Encode had researchers looking at the equivalent 1 percent of the genomes of more than 20 other mammals, and those results are forcing them to rethink the interplay between genetics and evolution.

The expectation was that many of the most active DNA sequences in humans would be prevalent in other mammals, too, because evolution tends to save and reuse what works best. But more than half were not found in other creatures, which suggests they may not be that important in people, either, said Ewan Birney of the European Bioinformatics Institute in Cambridge, England, a coordinator of the Encode effort.

"I think of them as gate-crashers at a party," Birney said. "They appeared by chance over evolutionary time . . . neither to the organism's benefit nor to its hindrance. That is quite an interesting shift in perspective for many biologists."