Two philosophers of science recently surveyed 500 geneticists to ask their opinion on whether 14 different sets of genetic information constituted a gene, or more than one gene. Fortunately, the bulk of the respondents felt able to answer the questions definitively. Less fortunately, their answers were inconsistent, with the sample often quite evenly split on the question of how many genes were actually present.

Sceptics might note a degree of unravelling here. Decades of discussion have left a rather widespread perception, embraced by the general public and the media, of the gene as a tightly defined entity that spells out an inescapable destiny filled with beauty and health or, more often, blemishes and disease.

Even among the medical profession and some scientists, a gene is a trusty and well-defined concept — a specific sequence of genetic information that, when converted into messenger RNA, encodes a protein. Following this line of thinking, all that medical researchers really need to do is link up those darned diseases with their underlying genes, and human biology would fall into place.

Among geneticists themselves, this notion has long been eclipsed. Where once scientists saw placid, lonely genes that mass-produce RNA transcripts, now they find a chaotic jumble of RNA generated from all over the genome and from outside conventional genes. They have little clue what this RNA is doing, and don't always know where one gene ends and the next begins (see page 399).

And anyway, forget about DNA, says a paper on page 469 of this issue; some RNA might also be ferrying information from one generation to the next.

The genetic code holds new allure — its four-letter sequence may have been documented but it contains deeper hidden ciphers.

For most geneticists, this complexity is a source of marvel and fascination — and employment. How dull their lives would be if, once the human genome had been sequenced, there were just genes and diseases to be linked, like one of those join-the-dots puzzles. The genetic code holds new allure — its four-letter sequence may have been documented but it contains deeper hidden ciphers, and geneticists relish the task of breaking them.

There remains a nagging concern, however, that some of these challenges will frustrate the hopes of earlier generations that the study of biology could reduce complex problems to a mechanistic understanding of the relationship between DNA and living things.

Indeed, the complexity of this relationship has got us to the point where geneticists find it hard to agree on an appropriate definition of a gene. They are also unsure whether genes themselves are worthy of the most attention, compared with other parts of the genome, or RNA or proteins, or the way they all interact together in different tissues. At the very least, a serious disconnect seems to have arisen between the real problems that geneticists are wrestling with, and the public understanding of what they do.

It falls on researchers to make sure that the gap doesn't grow too wide. That means conveying the complexity of the task more clearly and fighting the media's tendency to boil down complex investigation to the ‘discovery of the gene for something’. Geneticists should not be afraid of using new words, such as transcripts or loci, if these serve them better and more accurately. The public can cope with such distinctions.

After all, ‘gene’ is just a word, and dictionaries can be revised. Bring on the complexity — biology would be boring without it.