Features

BOSTON — Scientists are questioning the most surprising discovery from last winter’s deciphering of the human genetic code – the assertion that people have only about 30,000 genes, or roughly twice as many as the fruit fly. 

A new analysis suggests that number is too low, and the real total could be considerably bigger. However, researchers who came up with the original figure are sticking with it, at least for now. 

Scientists have long argued over how many genes it takes to build a human. Educated guesses have ranged up to 150,000. 

The issue seemed settled last February, when two competing scientific teams published the first detailed look at virtually the entire library of genetic information contained in every human cell. 

Both groups laid out the 3 billion bits of data that make up the code. Both used computers to distinguish the information that is genes from the look-alike filler. And both came up with roughly the same estimate: between 30,000 and 40,000 genes, with the best bet under 35,000. 

Some speculated that the relatively small number of human genes was good news, because it means less work to understand how they all work and perhaps translate that information into cures and treatments for various diseases. 

To many scientists, the fact that the two groups independently arrived at the same number made it believable. 

However, a team lead by Dr. Michael Cooke of the Genomics Institute of the Novartis Research Foundation in San Diego compared the two groups’ findings and found out that they had identified two quite different sets of genes, with only roughly a 50 percent overlap between them. 

The two groups agreed on the existence of about 17,000 genes. But about 25,000 more were found only by one group or the other. 

“It’s a jaw-dropper,” said Cooke, whose findings are published in Friday’s issue of the journal Cell. 

Just how many genes it takes to construct a human is unclear from the latest analysis. While Cooke believes 30,000 is too low, he estimates the total is probably not more than 60,000. 

For now, nobody knows how many genes were missed by both teams or how many of those identified by just one group truly are genes. 

One catalog of genes was compiled by Celera Genomics of Rockville, Md., the other by an international consortium headed by the National Human Genome Research Institute. 

Officials of both Celera and the consortium contend that most of the 25,000 genes found by just one group or the other will turn out to be phonies, so the final answer may still be somewhere around 30,000. 

“It’s way too simplistic to say you can add up the non-overlapping sets and get a bigger number,” said Celera President Craig Venter.  

“They’re probably bogus.” 

Dr. Francis Collins, head of the genome institute, agreed the total is unlikely to grow hugely. “It would not stun me if there turned out to be 50,000,” he said. “It would stun me greatly if there were 100,000.” 

Dr. Gerald Rubin, a fruit fly expert at the University of California at Berkeley, said some scientists suspected all along that the total number would turn out to be higher than 30,000. His guess: Humans will have 54,000 genes, or four times more than the 13,600 in the fruit fly Drosophila. 

Scientists are running a betting pool on what the total will be, and so far 165 have entered. (The cost of a bet rose from $1 to $5 after the release of February’s data.) Right now, their average guess is 61,710 genes. The winner will be chosen in 2003, by which time it is hoped the answer will be clear. 

Collins’ bet, made two years ago, was 48,011. “I’ll hold on,” he said. “I’m not completely retracting that.” 

On the Net: 

Cell: http://www.cell.com 

Betting pool: http://www.ensembl.org/Genesweep