Low number wins human gene sweepstake

The winner of the sweepstake that asked genetics researchers to place bets on how many genes would be found in the final draft of the human  genome sequence was announced last week. The most recent estimate, around 24,800 genes, is much lower than the 50-100,000 predicted some years ago. The closest guess in the GeneSweep contest, launched at the US Cold Spring Harbor Laboratory in 2001, was 25,947. The winner, gene sequencer Lee Rowen of the Institute for Systems Biology in Seattle, was one of only a handful of entrants who thought the human genome would contain less than 30,000 genes. Rowen took half of the $1200 pot, while runners-up Paul Dear of the UK's Medical Research Council and Olivier Jaillon of Genoscope in Evry, France, shared the remainder.

Although the human genetic code, or genome, is made up of around three billion base pairs of DNA, only around two per cent of this DNA sequence encodes genes. However, gene-predictor computer programs designed to locate these coding sequences are notoriously unreliable, reports last week's Nature. They often pick up 'pseudogenes' - defunct copies of real genes - and miss tiny genes, overlapping genes and small genes contained within larger ones. 'I'd say we don't know the true gene number for any organism - and certainly not for humans' said bioinformatics expert Phil Green of the University of Washington, Seattle. The current prediction of 24,847 human genes, made by the genetic database Ensembl, is unlikely to represent a final count, according to Jean Weissenbach, director of Genoscope.