Graduate students everywhere no doubt will be delighted by Mark Biggin's message: if you want to study transcription factors right, there's extra experiments waiting for you.

That's the conclusion senior author Biggin and his colleages at the University of California at Berkeley and Lawrence Berkeley Laboratory and Affymetrix Inc. reached in the February 2008 issue of PLoS Biology, after studying the relationship between transcription factor binding and changes in gene expression - which they found to be much weaker than they expected.

"There's a tendency to believe that everything you see is important," Biggin told BioWorld Today. "Our message is: a binding assay is a binding assay," whereas functional importance needs to be determined via separate studies.

Biggin said his team's long-term goal is to learn how to interpret the genetic code. Despite the fact that the entire human genome has been sequenced, the book of life is still for the most part undecipherable. "In the human case, we can interpret what 3 percent [of the genome] is doing - that's the protein-coding part," Biggin said.

As for the other 97 percent, it is clear that it doesn't just get transmitted because the cell's DNA replication machinery is obsessive-compulsive. But which sequences are important for what is still a mystery. "We can't read that code," he added.

Transcription factors are clearly a good place to start looking, because they control what cells make of their DNA. Given that every cell in the body contains the entire genome, if its use were not differentially controlled, "we'd all be big amorphous blobs of Jell-o," Biggin said.

The researchers studied six transcription factors that are made during drosophila development, and found that between them, they bound to thousands of sites in the drosophila genome.

But while binding was reproducibly observed at all those sites, the authors concluded that "many . . . poorly bound regions are not involved in early embryonic transcriptional regulation, and a significant proportion may be nonfunctional."

"The binding that looks to be most functional is the strongest," Biggin said. But "the total amount of what's nonfunctional versus what's functional is potentially quite large."

The authors based their conclusion on three main findings. Paradoxically, the putative nonfunctional sites are both too close and too far from genes. "Many poorly bound regions are in protein-coding sequences, [but] they are not (mostly) in the protein-coding sequences of genes transcribed in the early embryo," Biggin explained. Such sites also are less conserved than the strongest-binding sites.

Last year, the Encyclopedia Of DNA Elements or ENCODE Consortium also found a larger number of regulatory elements than expected in comparative genomics studies, and noted that many of them might possibly be nonfunctional. (See BioWorld Today, June 14, 2007.)

But Biggin said that while the ENCODE consortium advanced nonfunctionality as one of several possible explanations for the high number of binding sites they found in their study, his team has investigated predictors of functionality much more thoroughly.

Biggin's conclusions are somewhat sobering. "There's a strong tendency for people to believe that if they see something it must be functional, and to score things qualitatively," he said. In practice, that means setting a more or less arbitrary cutoff point and defining stronger binding as functional. But that, Biggin said, is not how cells work. Instead, "you need to look at it as a quantitative continuum. . . . The data can be, and should be, analyzed quantitatively."

Though Biggin and his colleagues studied transcription factors during development, he believes that the same basic principles hold for the way that transcription factors work in adult animals, in both normal and pathological situations.

Bacteria, which seem to be fundamentally different in their transcription factor binding sites, may be another story - but as far as animals are concerned, Biggin said that "we have certainly looked at many classes of transcription factors, and they all seem to fit the same pattern."