Tally of Human Genes Challenged Estimate May Be Higher Than Genome Project Predicted, Study Says
By Terence Chea Washington Post Staff Writer Friday, August 24, 2001; Page A10
Scientists are challenging one of the most widely reported findings to come out of the first detailed analysis of the human genome: that a relatively small number of genes are needed to make a person.
When researchers at Rockville-based Celera Genomics Corp. and the publicly funded Human Genome Project published their work in scientific journals in February, each group estimated the human genetic code was made up of about 30,000 genes, only about twice that of a worm or fly.
Now a study published in today's journal Cell argues that number is too low. Scientists at the Novartis Research Foundation's Genomics Institute in San Diego compared the two genome sequences and found that almost half the genes each group predicted did not overlap with the other group's genes.
If added together, they concluded, the overall tally of genes could be much higher than either group predicted.
"It underscores that Phase 2 of the genome project, which is finding all the genes, is still a work in progress," said Michael P. Cooke, coauthor of the article. "It will be a few years before we have a complete list of the genes that control human biology."
Scientists want to identify all the human genes to understand what causes disease and develop medicines to treat them. But only about 1 percent of the genome's approximately 3 billion chemical units make up genes, and the process of identifying those genes is far from an exact science.
Francis Collins, director of the National Human Genome Institute in Bethesda and chief of the international human genome project, wasn't surprised the Novartis group concluded there might be more than 30,000 genes. He said his group's paper had emphasized the uncertainties surrounding the preliminary estimate.
"It's clear that the methods we use to identify genes in long stretches of DNA are imperfect," Collins said. "People will be debating the precise number of genes for years to come, but we'll get a narrower band in the next few years."
Craig Venter, Celera's president and chief scientific officer, also agreed with the new study's conclusion. He said Celera's paper stated about 11,000 of the 39,000 genes his group predicted were hypothetical genes whose existence had not been confirmed.
"We still have a long way to go," said Venter, who predicted the final tally of genes would fall between 25,000 and 35,000. "We're still early in this field."
The gene tally is not the first challenge to the genome's analysis. Scientists have questioned another much talked-about finding: that hundreds of the genes were actually microbial genes left by bacteria that infected human ancestors millions of years ago.
That claim was challenged in May, when researchers at the Institute of Genomic Research in Rockville found the majority of those genes exist in other animals, suggesting they are just basic genes left over from evolution.
The 30,000-gene estimate came as a surprise because most scientists had believed there were about 100,000 human genes. They wondered how an organism as complex as a human being could be created with so few genes.
Many scientists have challenged the 30,000 figure. One article in the online journal Genome Biology estimated there were between 50,000 and 60,000 genes.
William A. Haseltine, chairman and chief executive of Human Genome Sciences Inc. in Rockville, has said his company has identified 90,000 genes and estimated that the total may reach as high as 120,000 genes. Most scientists dismiss that figure. Haseltine said the study underscores the "very preliminary nature and sloppy interpretation of the original data."
Although the two groups used complex mathematical formulas to make their predictions, Haseltine said company scientists isolated the genes themselves. "Ours is not speculation," he said. "Ours is based on actual physical observation."
John B. Hogenesch, the Novartis study's coauthor, said the new study demonstrates the value of having two genome maps, which allows for comparison.
"This really is a very complex problem, and we really don't have the answers yet," Hogenesch said. "Sometimes that's just as important as having the answer."