Charles Arthur: Why do we expect so much from science?

DNA is often called the blueprint of life but it would be better to call it a recipe

It isn't customary at birthdays for the guest of honour to be handing out the presents but that seems to be our attitude to DNA. Friday marks 50 years since the publication in Nature of a short note, called A Letter from Francis Crick and James Watson, in which they set out their idea of how DNA managed to squash so much information into such a tiny space.

And ever since then, we have been saying, "Yes, but what has DNA ever done for us?" Fifty years on, there is an undercurrent of impatience. Knowing that the difference of a single gene causes haemo- philia, or determines our eye colour, or creates a predisposition in certain families to breast cancer, has led to the expectation that once we have "sequenced" an organism's DNA into its mantra-like combinations of ATGGTTCCCTATG ..., then we know and can control everything about it.

Look at the expectations raised by the fact that a week ago Canadian scientists sequenced the Sars virus, after their team worked on it for six straight days, 24 hours a day. That this was feasible just a month after they began investigating is a tribute to the power of our sequencing systems. But the researchers who emerged blinking from their laboratories into the unaccustomed sunlight were entirely accurate to call it "the first step" to developing a vaccine – not least because other sequencing efforts have shown that Sars is actually mutating as it goes (entirely predictably for an infectious organism). Developing a vaccine will involve hitting a moving target. So far, only immune systems have shown much promise in doing that.

What about other "promises" that were allegedly made by DNA? What about gene therapy, whereby our flaky genes would be replaced by perfectly-tuned ones? Or that we would have our DNA "read" to see what diseases we were prone to and should guard against?

At least we can point to people who have been helped by gene therapy, although none have benefited so far who would not have been helped just as well by standard bone marrow transplants. Only with no suitable donors being available, gene therapy was the only way forward.

DNA, you might feel 50 years after we unravelled it, has let us down. But the reality is that simply unravelling it does not tell you how it works, any more than knowing the alphabet helps you to read a book.

The problem is that we expect too much of DNA. Or rather, that we expect too much of our limited understanding of DNA, which derives from the metaphors we use about it. DNA is often called "the blueprint of life", but as the writer Douglas Adams pointed out, it would be better to call it a recipe. Imagine that you were trying to describe how to make a fruit cake by writing the blueprint: currant here, surrounded by certain amount of air-filled cake mixture, and then more currants. It would be hellish. So how do we make fruit cakes? Not by blueprint. We use recipes – mix these things together, bake at a particular temperature for so long, and voila: if you've got the components right, you'll have currants distributed satisfactorily around your finished product.

DNA is something like that. Before the human genome was sequenced in its "rough draft" form, in June 2000, the expectation had been that humans would turn out to have more than 100,000 genes. To the surprise of many, the figure seems to be around 45,000 individual genes. To those used to thinking of DNA as a blueprint, that seemed like far too few; how could you build something as complex as a human with so few genes? After all, the puffer fish, the second vertebrate (after humans) to be sequenced has around 30,000 genes, while all it does is swim around and poison you if it's sliced up the wrong way for sushi. These are the sort of non-problems that arise when we think by metaphor.

Where the promises of DNA – promises that we attributed to it – have been fulfilled is in the diagnostic side. DNA fingerprinting has entered the language, both for crime and paternity. Medical tests for pathogens use DNA-based "probes" that can pick up the tiniest bit of a given chunk of nucleic acid. But on the therapeutic side, we're still a long way off.

That is not to say that these things are not going to happen. Fifty years from now, we will look back at our present expectations as hopelessly out of tune with our abilities. Perhaps, as with space travel, we will think that our hopes were out of kilter with what was actually needed, or reasonable.

The decoding of the human genome has been repeatedly compared with putting a man on the moon. That was 34 years ago, and we have not been back since – but we have got a lot better at putting satellites and rockets into orbit to look more closely at ourselves and our planet. Being able to read our DNA might be a similar step for mankind – but what is going to follow will be smaller, incremental, and ultimately more of much more use.