Scientists place bets on human genes

As the first draft of the entire human genetic code nears completion, speculation is increasing over the total number of genes in the human genome. Some delegates at the Cold Spring Harbour (CSH) genome meeting in the USA last week even started a sweep stake, taking bets at $1 a time. The guesses so far range from 27,462 to 153,478 - a spread that reflects the published estimates.

The uncertainty over the number of human genes stems from the different ways of counting them. Techniques that extrapolate from  genes identified in parts of the genome already sequenced (including the whole of chromosomes 21 and 22, relatively gene-poor and gene-rich respectively) give estimates of around 35 to 40,000 genes. A study published in the June issue of Nature Genetics, which compares the human genome to that of the puffer fish, predicts even fewer: between 28,000 and 34,000.

Other researchers have made estimates ranging from 100,000 - 140,0000, after using computers to analyse random fragments of genes isolated from different body tissues (Expressed Sequence Tags, or ESTs). But Tim Hubbard, who heads such a project at the UK's Sanger Centre, warns that this method may overestimate the number of genes. 'This is due to false positives and cases where multiple genes are annotated when there is really only one' he said.

Francis Collins, head of the US National Human Genome Research Institute, placed his bet on a modest 48,011 genes. He said that recent gene counts have been inflated, especially by genomics companies boasting that 'My list is bigger than your list.'

The speculating scientists will have to wait until the final version of the sequence becomes available, in around three years time, to find out who will pick up the cash prize. Those attending next year's CSH meeting will still be able to place bets, but they will cost $5, rising to $20 in 2002.