The draft human genome: Where are all the genes? 28/2/01. By Richard Gallagher and Carina Dennis The draft sequence of the human genome was found to contain many fewer genes than expected.

If a fly, a worm and a mustard weed can get by with 13 000, 18 000 and 26 000 genes, respectively, how many genes are needed to specify a human being? Previously, numbers up to and beyond 150 000 have been bandied about.

However, in something of a blow to our collective ego, it appears that there are only 30 000-40 000 human genes.

Proteins encoded by these genes can be grouped into families on the basis of their similarity to one another, and it turns out that we share most of the same protein families with worms, flies and plants - although the numbers of family members have expanded in humans.

This expansion is particularly evident in the genes that drive development by signalling between cells: humans have 30 fibroblast growth factors (fly and worm have two each) and 42 transforming growth factor-ßs (fly and worm have nine and six, respectively). Such differences are also apparent in the proteins of the immune system: humans have 765 genes with immunoglobulin subunits or 'domains' (such as those in antibodies), while the fly has 140, the worm 64, and the mustard weed and yeast have none at all.

The remaining additional genes are not primarily the result of invention of new types of protein in the vertebrate lineage only 7 per cent of identified protein and protein domain families are truly specific to vertebrates. Rather, new proteins come from reshuffling the number and order of protein domains, analogous to making new structures with the same Lego pieces.

If the increasing complexity of humans is not due to a dramatic rise in gene number, what might explain it? No single dominant property stands out. Instead, a mixed bag of features combine to greatly enhance innovation.

One example is alternative splicing of RNA: once RNA has been copied from a gene sequence, the non-coding sequences from the introns are sliced out to bring the coding sequences from the exons next to one another. By missing out an exon here and there, the splicing machinery can create novel products. Around 60 per cent of human genes have two or more alternatively spliced transcripts, compared, for example, with only 22 per cent in the worm.

Another factor is the lavish supply of proteins that switch genes on or off, such as the zinc finger family of proteins. This family has expanded independently in humans, yeast, fly and worm but humans still have twice as many zinc finger proteins as the fly and nearly five times as many as the worm. Meanwhile, proteins themselves can be modified, for example by enzymes snipping bits off them, or by the addition of sugars or fats to change their activity.

This builds into a picture of exquisite layers of control of genes and proteins, with genes being turned on and off, up and down, with extraordinary subtlety driving our development from fertilised egg to adult, and maintaining and repairing our bodies during the rigours of daily life.

Update: The analysis of the final sequence of the human genome was published in October 2004. Estimates of the number of human genes dropped again, to 20 000-25 000.

Image credit: Dan Salaman