Scientists Cast Bets on Human Genes; a Winner Will Be Picked in 2003 By NICHOLAS WADE How many human genes are there? Recent estimates have ranged from 60,000 to 140,000 and might be expected to narrow sharply as the human genome nears completion. In fact, the range has widened, suggesting that gene-counting remains at least as much art as science. Faced with estimates that are all over the lot, genome scientists have contrived a simpler way of settling the question, at least for the present. At a meeting this month at the Cold Spring Harbor Laboratory on Long Island they opened a sweepstakes for bets on the number of human genes. The winner will be chosen based on the most likely number decreed at the same annual meeting in 2003. The mean of the 228 bets so far cast is 62,598 genes, with a high of 200,000 and a low of 27,462. (Further details are on the Web at www .ensembl.org/genesweep.html.) By comparison, 19,099 genes are apparently required to run the C. elegans roundworm and 13,601 genes for the Drosophila fruit fly, the only two animals whose full genomes have so far been decoded. Given the range of estimates, the real number will have considerable bearing on how long it takes to understand the human genetic repertoire. Bets aside, three new estimates published last week in the journal Nature Genetics ranged from a high of 120,000 to a new low of 28,000 genes. Looking at the number of gene transcripts produced by living cells, researchers at the University of Washington calculated that the human genome contains a total of only 35,000 genes. An even lower estimate -- 28,000 to 34,000 genes -- comes from Dr. Jean Weissenbach of Genoscope in France. He and his colleagues tallied human genes by matching them to their counterparts in the puffer fish, where genes are easier to count. A third estimate, from Dr. John Quackenbush and colleagues at the Institute for Genomic Research, also based on gene transcripts, puts the number of genes at about 120,000. Until recently the low estimate has been one of 60,000 to 70,000 human genes, a figure arrived at in 1994 by Dr. J. Craig Venter, now president of Celera. The high estimate of 140,000 was issued last September by Dr. Randy Scott, president of Incyte, a commercial rival of Celera. Though Dr. Scott had a scientific basis for his claim, it did not seem entirely coincidental that his number was exactly twice Dr. Venter's highest estimate, as if to suggest twice the value in Incyte's gene set. Though Dr. Venter's number had long been the low end of the range, it suddenly seemed middle of the road earlier this month when a group of German and Japanese researchers calculated that there were only 40,000 human genes. Last week's estimates from the University of Washington and the French group support the idea that the number of human genes is far less than has long been supposed. These lower estimates, if true, could be considered serious threats to human pride and perceptions of self-worth. Can it really require only a third more genes to design and operate a person than a microscopic worm? By what sardonic sense of humor would evolution give humans the imagination to see themselves as the summit of creation yet make them with only twice the number of genes as required by a midget fly that feeds on rotten fruit? "We may have only twice as many genes as the worm but get a much greater increase in complexity from more complex interaction between the components," said Dr. Phil Green of the University of Washington, an author of one of the low-ball estimates. "That saves a little bit of our pride." For the record, Dr. Green's bet on the number of human genes is 35,000. Dr. Quackenbush's is 118,259.