he torrent of data spewing out of DNA sequencers in labs around the world is threatening to drown biologists even as it enables them to ask fundamental questions. No surprise, then, that a tool to tame the flood has been the most highly cited paper for two issues in a row. MEGA2 is the latest offspring of a software package that streamlines molecular evolutionary genetics analysis.

Masatoshi Nei, Professor of Biology at Pennsylvania State University and one of the founders of the field, tells Science Watch he was "surprised" at the popularity of the paper he co-authored with Sudhir Kumar and colleagues at the Arizona Biodesign Institute, Arizona State University. "We simply tried to produce a computer program package that can easily be used by investigators without spending much time," he says. "There are many other computer programs for phylogenetic analysis, but they are not easy to use for uninitiated researchers. For this reason, we also developed simple analytical approaches that can be understood intuitively."

Even without molecular data, figuring out the evolutionary relationship between two organisms is fraught with problems. If they share some particular characteristic, is that because they both inherited it from a common ancestor? Or is it because both have independently arrived at the same kind of solution to a common problem? If they are radically different, is that because they are not closely related, or because they are living radically different lives? Molecular data add several layers of complexity. DNA that codes for absolutely vital functions will be very constrained and will not change much; it thus offers little information about relationships. And DNA that does not code will be free to mutate, but may then mutate back. Then there are the difficulties created by insertions and deletions, rearrangements of larger chunks of sequence, and so on.

A variety of methods have been devised to deal with these and other complexities, and, as Nei points out, there are a huge number of computer programs designed to help researchers use molecular information to explore evolutionary history. Kumar, now Director of the Center for Evolutionary Functional Genomics at Arizona State University, started working on MEGA in 1991 while a Ph.D. student in Nei’s lab. He says that one of the great strengths of MEGA2 is the way it handles large datasets. The difficulty arises because the number of possible evolutionary trees goes up exponentially as the number of branches increases. "Finding the optimal tree is impossible to do," Kumar tells Science Watch, adding that "you cannot exhaustively compute" all the options.

Nei had helped to devise a technique called the neighbor-joining method, which gave a very good approximation to the best tree (N. Saitou, M. Nei, "The neighbor-joining method: a new method for reconstructing phylogenetic trees," Molecular Biology and Evolution, 4[4]: 406-425, 1987), which is something of a citation champion itself, by now having logged more than 10,000 cites. "With MEGA you can do a neighbor-joining tree in less than a minute on an ordinary PC," Kumar says. The results bear that out. MEGA2 has been used with huge datasets, for example a recent analysis of 1,060 sequences of human mitochondrial DNA.

Nei has recently used MEGA to build phylogenetic trees for hundreds of MADS-box genes (MADS-box genes are highly conserved sequences that are widely distributed and that seem to control many important aspects of development; some biologists believe they will prove as important as homoeobox genes). The genes were identified from the Arabidopsis and rice sequences but their function is currently unknown. "Our analysis," Nei tells Science Watch, "identified several groups of genes whose functions are likely to be similar within groups." (J. Nam, et al., "Type I MADS-box genes have experienced faster birth-and-death evolution than type II MADS-box genes in angiosperms," PNAS, 101[7]: 1910-15, 17 February 2004). Experimenters are now following up on suggestions that emerged from the phylogenetic trees to work out what these genes do. And that could have profound consequences because other MADS-box genes are known to be involved in the timing of flowering. Manipulating those genes could offer a route to improving the yield of rice and other cereals.

This, in essence, is how both Kumar and Nei see the usefulness of MEGA2. "Researchers can obtain the results quickly and spend more time for thinking about the biological meanings of the results," Nei says. "MEGA2 is a biological tool for biologists, not a bioinformatics tool for bioinformaticians," notes Kumar. "The focus has been on the bench scientists all along, to enable them to discover novel patterns through comparative sequence analysis."

As more and more data become available, so the task of making sense of it becomes harder and harder. Are more powerful computers and cleverer programs the answer? Not for Nei. "I do not think high-speed computers can solve major problems in evolutionary biology," he tells Science Watch. "The most important equipment is the human brain. Only human brains can extract important scientific principles from vast amounts of DNA sequences. However, computers are certainly useful for sorting out important factors."