When the Kentucky Derby winner crosses the finish line in front of 160,000 roaring spectators on May 5, there's a good chance it will have two copies of a gene that makes a horse a sprinter.

The so-called speed gene, which several laboratories say determines whether a horse prefers a short sprint, a marathon or something in between, is just one of the genetic markers identified in the search for the roots of elite performance in thoroughbreds. Now the race is on among five or six commercial laboratories to convince thoroughbred breeders and buyers that testing for this gene and other markers is the road to the Triple Crown. In the meantime, the geneticists behind these companies scramble to lay claim to the best markers for athletic traits. Major thoroughbred farms are signing up horses for testing, even though some say they're not sure what the results mean.

"We don't know what to make of it," says Elliott Walden, president, CEO and racing manager of Winstar Farms in Versailles, Ky. Winstar, the 685-hectare birthplace of 2010 Kentucky Derby winner Super Saver, is dabbling in genetic testing. "We don't know how to evaluate the information. We're still figuring it out."

He's not the only one puzzling over these tests, which start at around $500 per horse.

A presumed genetic advantage
Equix Biomechanics is a company driven by data. The Lexington, Ky., concern advises on thoroughbred buying and breeding after gathering a mountain of data including 36 separate measurements of bone and muscle groups and a computer analysis of the horse's gait. Operated by Equix president, J. Todd Stewart, who in his past life performed statistical arbitrage for hedge funds, and owned by another numbers guy, Gary Knapp, who also came from a career in finance, the firm feeds its many measurements into a mathematical model that spits out, theoretically, the perfect horse—one that moves quickly, efficiently and—for those with Triple Crown aspirations—with enough endurance to win the longest of the three races, the 2.4-kilometer Belmont Stakes. Big Brown, winner of the 2008 Kentucky Derby and Preakness Stakes, was bred by the numbers in this formula on Knapp's Monticule Farm in Lexington. Now, the company is adding gene profiles to its analysis.

Knapp recently had some of his own horses profiled. "I wasn't sure of the actual relevance," he says. "There are just so many variables impinging on the relative success of a particular runner that it's difficult to take one particular variable and say, 'That's the key.'"

Alongside this growing interest in the emerging genetic field are beloved traditions of an earlier century—the heavy reliance on pedigree to determine racing quality. Pedigree remains the major predictor of a horse's sale price, says geneticist Matthew Binns, of Midway, Ky.–based The Genetic Edge and former professor of genetics at the University of London's Royal Veterinary College. Unfortunately, it is a less capable predictor of performance. People view pedigree as a surrogate for genetics, Binns says, "but there's a very poor correlation between pedigree and DNA data."

Convincing the thoroughbred industry of pedigree's limits will take a revolution. Buyers routinely look back to a horse's 32 great, great, great grandparents when deciding which animal to buy. Binns points out that, on average, a great, great, great grandparent can contribute only about 3 percent of its genes to any individual. But even data-loving Knapp is in the sway of pedigree, extending his own analyses back not just five generations, but to the 128 ancestors in generation seven.

Pedigree versus genetic testing
Trading the pedigree tradition for genetic testing could save thoroughbred buyers money, the gene profiling companies contend. Emmeline Hill, the discoverer of the "speed gene," leads the Equine Exercise Genomics Group at University College Dublin in Ireland. She is also chairman of the commercial horse genetics laboratory, Equinome, Ltd., which she co-founded. She says her research shows buyers consistently overpay for horses that perform poorly. She evaluated the sale prices of 200 yearlings and compared the outlays with the animals' on-track performance.

"When we looked at total earnings, we found buyers were paying far too much for poor quality horses," Hill says. The better predictor for earnings, she adds, was how the horses ranked on Equinome's report card, which assigns horses to a performance class based on how its genome compares with those of elite horses in her database, with class 1 being the best and 4 the worst. Her message: Not even the experts can distinguish a 1 from a 4 if they don't know the genes. "When we did the statistics, we found class is a more accurate predictor of total earnings than yearling sale price," Hill says. "Buyers were paying just as much for class 4 as for class 1. This is a bit of a fright, really."

The classic case of all this horse sense gone wrong is the story of The Green Monkey, a handsome bay colt with top-notch bloodlines that went to auction in 2006. The colt further impressed shoppers with a fast breeze—a breeze is a test run two-year-olds make before auction. Bidding began at $2 million that day in Florida. In little more than eight minutes, the horse sold for $16 million, the highest price ever paid for a thoroughbred in public auction.

It ran only four times and never won a race.

Genes and maps
When it comes to published academic research, Hill may be champion among the geneticists engaged in the search for the secrets of equine athletic potential.

In addition to publishing the discovery of the "speed gene"—actually a gene that controls the expression of the muscle growth factor, myostatin, in a January 2010 Public Library of Science paper—she has published research on several more genes and is seeking patents for some, including myostatin (MSTN).

Despite its nickname, the gene doesn't really play a role in how fast a horse runs, but rather how far it likes to run. Horses homozygous (having two copies) of one version (allele) of the gene—the so-called CC horses—are sprinters, she says. The CC means the horse has two copies of the DNA base cytosine at a particular place on the myostatin site. Those homozygous for the DNA base, thymine, at that site (the TT horses) are endurance performers. And those with one C and one T can show up on either side of the spectrum.

Among the other genes she says play a role in thoroughbred performance are peroxisome proliferator-activated receptor-γ coactivator-1 alpha (PGC-1α), which is involved in skeletal muscle adaptation to exercise; pyruvate dehydrogenase kinase, isozyme 4 (PDK4), involved in glucose regulation, expression of which increased almost fourfold after sprints; and cytochrome c oxidase subunit 4 isoform 2 (COX 4I2), involved in respiration. Hill showed a single-nucleotide polymorphism—placed in the genome where a single nucleotide is altered—in this gene is "strongly associated" with elite sprint racing performance.

Hill's assessment of thoroughbred performance is heavily dependent on what myostatin allele it starts with, and she has often looks for performance genes in the light of a horse's myostatin status. For instance, a study in her lab suggested that PDK4, which contributes to glucose regulation, may play a role in thoroughbred performance. When a follow-up study failed to replicate the finding, she decided to look again, but this time she sorted the horses according to their myostatin allele. Her finding: PDK4 alleles affected the performance of only those horses with the long-distance TT myostatin allele.

The approach is unique to her lab. In terms of uncovering specific genes and publishing on them, she nearly has the field to herself. Few other thoroughbred geneticists have expressed interest in identifying individual genes.

Genome-wide studies
There is a reason for that. Finding a specific gene responsible for racing performance "takes a lot of work and a lot of time," says Steven Tammariello, the lead geneticist for Lexington-based Performance Genetics. "Ultimately, it may not be that useful." It also costs a lot of money.

Instead, companies rely on genome-wide association studies, which scan for patterns and similarities. For instance, Tammariello looked at 65,127 single-nucleotide polymorphisms (SNP) in hundreds of horses, half of which were winners of elite races. In that search, he found 12 SNPs that associated with the best horses. These SNPs are the basis for any tests Performance Genetics conducts; it looks to see what alleles each tested horse possesses at these 12 locations. The more the horse has in common with a champion, the better it scores.

As for the "speed gene" Tammariello, also an associate professor of biology at Binghamton University in New York State, dismisses the prominence Hill accords it. "We found three other places that are also very important to distinguishing a sprinter from a route runner." And forget about it as an indicator for speed. "I owned two sprinters. I know they had the CC polymorphism, and they were terrible. It has nothing to do with the class of the horse," he says.

Dav Doodnauth, an MD and CEO of LifeLine Labs in Lexington, also uses genome-wide association studies to produce his "Pegasus Profiles." The internal medicine specialist says myostatin plays only a minor role in his analysis. "I don't think there is necessarily one area of the genome that's going to tell you everything about performance," Doodnauth says. Human athletic performance involves interactions of some 220 genes; the situation is likely to be the same in horses. "Myostatin markers are just a handful of the makers that we look at," he notes.

The Genetic Edge's Binns, who also published a research paper laying claim to the myostatin finding about a year after Hill did, says half of the 10 or so Kentucky Derby winners his company tested had two copies of the sprint version of the myostatin gene. One horse had two copies of the endurance version of the gene, and the rest had one copy of each. (He wouldn't say which winner had which combination.) His analysis: The "speed gene" isn't the sole determinant of Derby victory. "I think you get into difficult water if you base your breeding program on one gene," he says.

Nor will studying a horse's build and posture reveal what genes tell, Binns says. "You can't see these markers," he adds. "The best looker in the world cannot see these markers."

Other "silver bullets"
If the notion of a genome laid bare is enticing to some breeders and buyers, it's worrisome to those who have watched other technologies become the newest new thing.

"Everybody's looking for the silver bullet," says Kerry Cauthen, managing partner of the consignor Four Star Sales in Lexington—a job he says is somewhere between a matchmaker for horse buyers and sellers and the person who authenticates paintings at Sotheby's. He's seen such purported silver bullets before. Two examples: sometime in the 1990s, shoppers began demanding x-rays for any horse they considered buying. Whereas this was a useful tool in some cases, in others it led buyers to eliminate animals unnecessarily from consideration, Cauthen says. For instance, when an x-ray showed osteochondritis, or a split in the cartilage, in the hock—on a horse's hind leg, analogous to our ankle—buyers would lose interest, and good horses lost value. "It just tanked a sale. This horse is now worth a quarter or half of what he had been," Cauthen says. But osteochondritis is common in young animals, and often has no impact on a horse's training and performance, as later evidence showed.

Buyers similarly became enamored of endoscopic examinations of a horse's upper respiratory tract to look for abnormalities of the pharynx and larynx that can hamper airflow. Anything less than a top scope rating scares buyers away, Cauthen says. Then Scott Pierce, a veterinarian at Rood and Riddle Equine Hospital in Lexington, studied the correlation between endoscopy scores and subsequent racing performance in more than 800 horses. In a paper published in the 2001 American Association of Equine Practitioners's Proceedings, Pierce showed that only the lowest endoscopy scores hurt performance, and horses with mid-range scores were no more likely to perform poorly than horses with top scores.

Proprietary data
For Cauthen, evaluating any testing, including genetics, comes down to data. He'd like to see some. "Show me your results. Show me you were able to prospectively select winners based on your data," he says.

Most of the companies make no bones about keeping the performance data Cauthen looks for to themselves. The Genetic Edge, owned by Equine Analysis Systems (EAS), adds genetic testing to a series of other performance measures including motion analysis, metabolic profile testing, breathing tests and heart testing. Taken together, veterinarian David Lambert, EAS's managing owner, says these tests led his clients to breed or buy 27 Kentucky Derby runners, including two Derby winners, two Horse of the Year winners, two Dubai World Cup winners, plus more than 60 graded stakes winners.

But it's impossible to test the claims of champion horses selected by EquineAnalysis System Inc.. His clients are unidentified. And no one talks about the horses that didn't win.

Doug Cauthen, a thoroughbred manager who helps people make decisions about their horses, has used genetic testing since his days as farm president and CEO at WinStar. Cauthen is brother of consignor Kerry Cauthen and of Steve Cauthen, who in 1978 rode Affirmed to win the Triple Crown.

For Doug Cauthen, employing genetic testing is simply a matter of keeping on top of emerging technology. "I think there's something to it," Cauthen says. "It's a tool that will get a lot better over time. You should be open to it. It can't by any stretch of the imagination be the primary decision, because these are animals, and the primary decision has to be how they look and act. But it could help, and it's evolving."