http://www.economist.com/editorial/freeforall/current/st2340.html -- Economist SCIENCE AND TECHNOLOGY Drowning in data Like so many others, biologists are confronted by a tidal wave of information. Unfortunately, few of them know how to swim ONCE upon a time, biology was simple. Its practitioners cultivated things in Petri dishes and flowerpots, or studied them through fieldglasses. They might count them, measure their lengths, or even weigh them. But the numbers-and the crunching needed to interpret those numbers-rarely taxed their mathematical skills beyond a level that they would have learned at school. That is, however, changing fast. Biological data are flooding in at an unprecedented rate. The amount of information stored, for example, in the international repository of genetic sequences known as GenBank is doubling every 14 months. As a result, many of the challenges in biology, from gene analysis to drug discovery, have actually become challenges in computing. Indeed, the process of change is so rapid that some of the subject's potentates are afraid that progress may grind to a halt unless a huge injection of numeracy takes place pretty soon. The mightiest of those potentates inhabit America's National Institutes of Health (NIH)-the body responsible for disbursing the lion's share of federal money available for biomedical research. And earlier this month the NIH issued a report that talked of "the alarming gap between the need for computation in biology and the skills and resources available to meet that need" and recommended spending up to $160m on rectifying matters through a network of biocomputing centres across the country. An embarrassment of riches The main reason for this shotgun marriage with information technology is that biology has belatedly realised that it is, itself, an information technology-even though the technologist is natural selection rather than Bill Gates. An organism's physiology and behaviour are dictated largely by its genes. And those genes are merely repositories of information written in a surprisingly similar manner to the one that computer scientists have devised for the storage and transmission of other information-that is, digitally. There are superficial differences, of course. The genetic code has four elements (the four so-called bases, sometimes referred to as its letters), rather than the two of a binary coding system. And the bases are grouped together in threes, known as codons, rather than in the eight-bit bytes of computing. But the similarities are more striking, so the subject is suddenly lending itself to a serious amount of computerisation. At the same time, there has been rapid progress in the machines that supply the raw material-the sequences of genetic letters and codons in chromosomes. A single high-throughput gene-sequencing machine can now read hundreds of thousands of bases per day; and newer technologies, such as "gene chips", should make the analysis even faster. That will produce even more data that have to be stored and annotated for subsequent study. And even for those who do not work directly on the genes themselves, similar technological changes are appearing. Robotic screening machines, for example, in which hundreds of compounds in tiny wells are tested to see if they react with a particular biological target, can analyse thousands of compounds in a day. The result is a mind-boggling amount of information. According to Anthony Kerlavage of Celera, a company formed last year with the intention of sequencing the entire human genome using private money (and beating government-financed projects in the process), a genetics laboratory can easily produce 100 gigabytes of data a day-that is about 20,000 times the volume of data in the complete works of Shakespeare or J. S. Bach. The analysis of such data poses problems beyond mere volume control. Having sequenced a particular piece of DNA, for example, it is useful to compare it with a central database (such as GenBank) of existing sequences to see what it resembles. But this requires more than just a straightforward database search. The program involved must know what constitutes a biologically meaningful resemblance, and it must also be able to deal with the errors that inevitably creep into the sequencing process. As a result, devising new search algorithms requires extensive knowledge of computing theory, together with a keen biological intuition. And there's the rub. The real problem about the growing quantification of biology is not the change in the subject but the lack of change in its practitioners. For a sudden inpouring of data is not unique to biology. Astronomers, who once squinted over photographic plates, now deal with squillions of bits of data from automatic sky surveys. Meteorologists no longer use seaweed; instead, they prefer supercomputers. Particle physicists would not have the first idea of what was going on in their machines if the results of their experiments were not processed automatically. Yet none of those fields seems to be suffering unduly from information overload because the physical sciences are founded on number-crunching. Astronomers, for example, have been using rooms full of computers ever since the days when the word "computer" referred to a skilled mathematician. And some of the first electronic computers were devised specifically for use by physicists working on the development of atomic weapons. Many biologists, however, avoided the fields of astronomy, meteorology or particle physics precisely because they have, in the delicately chosen words of Sylvia Spengler of the Centre for Bioinformatics and Computational Genomics at the Lawrence Berkeley National Laboratory in California, "some problem with mathematics". The result, according to Larry Hunter, president of the International Society of Computational Biology, is that there is a desperate shortage of specialists capable of developing the computational tools that biologists need. What is required, he says, is "a genuinely new kind of scientist" who is trained in both computer science and biology. Worryingly, however, the demand for computational biologists is such that the very academics needed to teach interdisciplinary courses that might plug the gap are going into industry, where their skills are more highly remunerated. Some physical scientists used to accuse innumerate biologists of "physics envy". Partly, the accusation was that they secretly envied a numerical rigour to which they could not possibly aspire. Partly, it was that physicists got all the money. Now, however, it is the biologists' budgets that are growing. But there is a price. As biology becomes numerically rigorous, its practitioners have no choice but to do the same.