(FORTUNE Magazine) – First, a confession: Weeks ago I grew weary of the relentless roll of journalistic drums about the imminent decoding of the human genome. Sure, it's biology's moon shot. True, it will pave the way for gene-based cures for everything from cancer to cellulite. And, yeah, it's incredible how fast Celera Genomics--led by surfer-turned-Vietnam-war-medic-turned-yacht- racing-biowizard Craig Venter--has closed in on the milestone, vying with the publicly funded Human Genome Project for bragging rights to the first complete copy of the book of life.

But enough already. By the time the genome hullabaloo inflated Celera's market cap to $14 billion this spring, I was ready to consign the story to the annals of tulipomania and move on. Then I encountered a startling strain of brain-damaged flies that cured my genome fatigue and brought home what Celera's strategy is really all about.

Contrary to what investors seemed to think as they pumped the stock price of Celera and other genomics companies up and down over the past six months, decoding the genome is not about getting dibs on a pot of gold. Most evidently believed that Celera's business plan hinged on winning the race to spell out our DNA's "raw" sequence, the order of the three billion or so chemical units that make up the genome. If that were true, Celera would be a dumb bet. One reason is that the sequencing data are mostly a readout of "junk" DNA, the genomic equivalent of plastic packing peanuts. Commercially valuable (and patentable) genes, which determine everything from gender to disease susceptibility, exist as fragments embedded in the junk. It will take decades to winnow them all out, determine their functions, win key patents on them, and develop revolutionary gene-based medicines. Celera, in Rockville, Md., will be just one of many players in this huge enterprise, along with genomics leaders such as Millennium Pharmaceuticals and Human Genome Sciences.

This long-term game will unfold on a level playing field as far as access to the raw human DNA sequence goes--the Human Genome Project has already put nearly 90% of it on a free Website (www.ncbi.nlm.nih.gov/genome/seq).

Thus, Celera's vaunted first copy of the genome, scheduled for completion in June, won't generate any more intellectual property by itself than Neil Armstrong did when he set foot on the moon. Yet anticipation of this symbolic event drove Celera's share price from about $30 to $276 a few months ago. And when President Clinton and British Prime Minister Tony Blair announced in mid-March that they support free availability of the raw sequence data--inciting investor panic and erasing nearly $7 billion of Celera's market value by April--they were merely underscoring the genome project's long-standing public-release policy, not attacking the validity of gene patents.

But Celera isn't simply racing to pull off a historic publicity stunt. It aims to become a leading seller of genomic information to drug companies. To do that, it must show that its main asset--rows and rows of DNA-sequencing machines, coupled with supercomputers to analyze the data they produce--can give customers an advantage in the competition to identify disease-related genes and analyze how they work. One of Celera's selling points will be its ability to compare individuals' DNA to find genetic variations underlying idiosyncratic susceptibility to diseases. Perhaps its most exciting near-term edge, though, will spring from its leadership in a hot new field called comparative genomics.

Comparative genomics is the study of similar genes in different species. Celera is setting itself up to lead the field commercially by decoding the genomes of one creature after another over the next few years. The benefits of this strategy began accruing a few months ago, when Celera helped academic scientists decode the genome of the fruit fly. A research team including none other than Craig Venter identified in one fell swoop 177 fly genes that are strikingly similar to human genes linked with various diseases, from cancer to diabetes. By tweaking these genes in flies, researchers hope to learn lots about how they work--precious clues for developing medicines to correct the malfunctions that make people sick.

For a fascinating glimpse of comparative genomics at work, consider the ailing flies I met at a Harvard Medical School lab. They were produced by Mel Feany, a gifted young pathologist with a knack for twisting fruit flies' brains in interesting ways. She mutated a gene linked to Parkinson's disease in order to give the insects a progressive illness that looks remarkably similar to the human disorder.

Around the middle of their 60-day lives, her flies begin to lose their climbing ability, a sign of the kind of motor deterioration that robs Parkinson's patients of mobility. Neurons in certain parts of the flies' brains develop protein globs closely resembling Lewy bodies, telltale abnormalities that appear in Parkinson's patients' neurons. The insects also become hypersensitive to tremor-inducing drugs--their legs shake at doses that don't faze normal flies.

But what's most remarkable about Feany's study--and the part that best highlights the power of comparative genomics--is that she implanted a mutated human gene to cause these effects. The flies, in a sense, suffer from a human disease, not a fly one. The line of reasoning that led Feany and a collaborator, professor Welcome Bender, to create such "humanized" flies illustrates why comparative genomics is taking off.

The scientists began with the fact that many genes, especially those underlying basic cellular mechanisms, haven't changed very much while passing down from species to species over the evolutionary eons. At the level of their molecular building blocks, fly and human brains seem two chips off a very ancient block.

Dopamine molecules, which transmit signals among the kind of neurons destroyed by Parkinson's disease, as well as other brain cells, are exactly the same in flies and people, notes Feany. And like humans, fruit flies are semi-educable--memory mechanisms resembling some of ours enable them to remember and avoid odors they're exposed to before getting an electric shock. (Feany is well versed in flies' remembrance of things past: She earlier isolated the "amnesiac" gene, which, when mutated, causes flies to rapidly forget the shocking little things they learn.)

Nature's genetic conservatism is a boon to researchers trying to elucidate gene functions. To get a general idea of what a newly found human gene does, for example, they can mutate structurally similar genes in lower animals and watch what goes wrong--a strategy akin to punching a hole in a Model T's gas tank to get a basic idea of what the analogous structure does in a business jet. Mutating selected genes in creatures like flies and mice sometimes produces humanlike diseases in the animals, revealing the genetic underpinnings of the illnesses. Feany's flies represent a variation on this theme.

The messed-up gene she implanted was identified three years ago as the cause of a rare inherited form of Parkinson's disease in an Italian family. Feany believes this "alpha-synuclein" gene doesn't merely stop working when mutated. Instead, she says, it probably triggers the production of toxic proteins that gradually kill certain neurons. Given the deep similarities between fly and human neurons, she guessed that implanting mutated versions of this gene in flies might trigger a similar toxic process.

To test the idea, she injected fly embryos with DNA containing mutated alpha-synuclein genes and subjected the tweaked insects to a climbing test as they aged. Demonstrating the results, she taps two bottles crawling with flies on a countertop, causing them to fall in heaps at the bottom. One bottle contains middle-aged Parkinson's flies, the other middle-aged normal ones. Within 20 seconds many of the normal flies have crawled to the top--flies instinctively avoid being at the bottom of anything. Meanwhile, only a couple of their Parkinson's kin have clambered up. Several have toppled over and seem paralyzed.

Now Feany is teasing out fly genes that interact with the mutated one as it wrecks brain cells--a straightforward task, thanks to the well-stocked toolbox that fly geneticists have developed over the years. With such genetic co-conspirators in hand, she'll be able to use computer programs to identify similar human genes in a matter of hours, if they exist, as well as the current best guesses about their roles in the body. That could speed drug companies' quest for better gene-based treatments. Before comparative genomics began to blossom a few years ago, it typically took years to pinpoint a key disease gene.

With Harvard neurologist Peter Lansbury, Feany also plans to use her flies to assess prototype drugs for Parkinson's disease--since they're humanized, the flies should make very good test subjects. "I've had many, many calls from drug companies" since the study was reported in the March 23 issue of Nature, she says.

Cross-referencing the recently decoded fly genome with the soon-to-be-completed human one promises scores of similar models for studying human diseases. A team at the University of California at Los Angeles has created a condition in flies akin to the brain deterioration of Huntington's disease by implanting a human gene. Feany believes a similar fly model of Alzheimer's disease will soon emerge. Exelixis, a comparative-genomics pioneer in South San Francisco, is studying cancer using flies implanted with human p53 genes, whose malfunctioning helps trigger many kinds of tumors, says chief scientific officer Geoffrey Duyk.

Comparative genomics also helps researchers spot gene fragments hidden in raw DNA sequences, a job akin to parsing words that are broken into pieces and scattered through long strings of nonsense letters. That's because junk DNA changes faster than genes do--survival of the fittest, working at the level of molecules, tends to preserve well-honed genes while allowing the do-nothing junk to mutate randomly. Thus, genes can be found by scanning raw DNA from, say, mice and humans for pieces that haven't changed much over the past 50 million years.

All these applications make the ability to do cross-species comparisons essential for major genomics companies. Genetic-information vendor Incyte Genomics, for instance, is adding data on rat, mouse, monkey, and beagle genes to complement its human databases. But Celera may pull ahead. By combining publicly available data with its own, it expects next year to have in hand the full yeast, roundworm, fruit fly, mouse, and human genomes. On its tentative to-do list: rat, dog, and chimpanzee genomes as well as those of major crop plants.

Rounding up all these genomes would give Venter & Co. boasting rights, at least for a while, to a unique resource for probing genetic heirlooms that go back a billion years. That may not justify a $14 billion market value. But even after the current genomania dissipates, it will make Celera well worth watching.