News this Week

Science  22 Aug 2003:
Vol. 301, Issue 5636, pp. 1026
1. HIGH-ENERGY PHYSICS

Showdown Looms for Reigning Particle-Physics Theory

1. Charles Seife

BATAVIA, ILLINOIS—Will a rogue penguin demolish the Standard Model of particle physics? Scientists at a meeting* here hope so. Last week, researchers from Japan told startled colleagues that a peculiar type of particle alchemy known as a “penguin” may have revealed holes in the reigning theory. Although it's too early to draw a firm conclusion, physicists believe that within a year the penguin either will be tamed or will have trampled the Standard Model firmly under its webbed feet.

“There's a very large discrepancy between what we measure and what the Standard Model predicts,” says Stephen Olsen, a physicist at the University of Hawaii, Manoa, and member of the Belle team at Japan's KEK B factory in Tsukuba. “If true, it would be the first evidence of the breakdown of the Standard Model.”

The Belle experiment investigates the decay of a middleweight particle called a B meson, which contains a “bottom” quark. Once in a long while, the bottom quark spontaneously decays into two very massive objects: a “top” quark, the heaviest of the fundamental components of matter, and a W boson, a massive particle that carries the “weak” force. The W and top quark are transformed into three “strange” quarks; when diagrammed with a standard physics shorthand, this transformation looks vaguely like a penguin (see figure). The Standard Model makes firm predictions about how often and in which contexts these penguins appear.

Since it betan operating 4 years ago, the KEK B factory, which smashes electrons into their antimatter counterparts, has created more than 150 million B mesons. By studying how these B mesons decay, particle physicists in Japan and at a similar facility at the Stanford Linear Accelerator Center (SLAC) in California measure numbers that dictate how quarks transform into one another, a key aspect of the Standard Model of particle physics. One of those pillars is the expression sin 2β, which plays a role in the fundamental asymmetry between matter and antimatter. In 2001, the BaBar experiment at SLAC announced evidence that sin 2β was lower than the Standard Model predicted, but measurements from Belle seemed to bring the value back in line (Science, 23 February 2001, p. 1471). No longer.

To the audience's surprise, Belle collaborator Thomas Browder of the University of Hawaii, Manoa, announced that Belle's latest value of sin 2β is much lower even than BaBar's value. In statistical terms, the measurement is a 3.5-sigma departure from the Standard Model. “A lot of people are not just stunned by this but shocked,” says Ikaros Bigi of the University of Notre Dame, Indiana.

The difference might indicate that particles unaccounted for in the Standard Model subtly alter the penguin for this kind of B-meson decay, changing the effective value of sin 2β. That would explain why many measurements based on other decays agree on a value of sin 2β that is different from the ones Belle and BaBar see. Before declaring that physics has pushed beyond the Standard Model, however, scientists must collect more B-meson decays to make sure that the effect is real and determine why Belle's result is even more extreme than BaBar's.

The answers should become much clearer by this time next year, says Marcello Giorgi, a physicist at INFN in Pisa, Italy, and member of the BaBar group. “We are living in a very exciting period,” Giorgi says. “By next summer, there could be an announcement of new physics.”

• *Lepton-Photon 2003, 11–15 August.

2. HIGH-ENERGY PHYSICS

Visa Delay Foils Chinese Delegation

1. Charles Seife,
2. Ding Yimin*
1. Ding Yimin writes for China Features in Beijing.

BATAVIA, ILLINOIS—Chen Hesheng and his colleagues at the Institute of High-Energy Physics (IHEP) in Beijing had high hopes of attending last week's particle physics conference here (see related story). But as other lab directors listened to presentations of the latest findings, the Chinese scientists sat home and fumed. Although U.S. embassy officials interviewed Chen on 20 May in Beijing, they didn't approve his visa until 14 August—2 days before the conference closed and not nearly in time for him to make the trip.

Chen is only the latest foreign scientist to experience the heightened border security after the 11 September terrorist attacks. Although nobody disputes the need for greater vigilance, there is also consensus that the current environment is damaging international collaborations. “It is a problem for U.S. particle physics, and it's a problem at all universities,” says Michael Witherell, director of the Fermi National Accelerator Laboratory here.

IHEP has asked the International Union of Pure and Applied Physics, which sponsored the meeting, to avoid scheduling any more gatherings in the United States until the visa process is speeded up. The union's particle physics committee is already preparing a formal letter protesting Chen's treatment. “Normally, there are 30 or 40 Russians and about 20 Chinese at this conference,” says Stanford University's Vera Luth, chair of the committee. “This year there was one Chinese and four or five Russians.”

Some scientists find irony in the location of the latest barriers to the free flow of information. “In the past, it was always our countries that made it difficult to get a visa,” notes Michal Turala of the Institute of Nuclear Physics in Krakow, Poland.

3. ANTIBIOTIC RESISTANCE

WHO Advises Kicking the Livestock Antibiotic Habit

1. Dan Ferber

Eliminating the routine use of antibiotics in livestock reduces human health risks without significantly harming animal health or farmers' incomes, according to a World Health Organization (WHO) report released last week. The report adds to a growing momentum to end the use of antibiotics to promote growth in farm animals. The European Union (E.U.) has ordered member countries to end the controversial practice by 2006, and fast-food giant McDonald's is pressing its meat suppliers to cut back on antibiotics as well.

The WHO review of experiences in Denmark, which has phased out the practice, is the most comprehensive assessment to date of the effects of banning antimicrobial growth promoters, says Peter Braam, WHO's project leader on food-borne diseases, who coordinated the effort. “Under conditions similar to those in Denmark,” the panel wrote, “the use of antimicrobials for the sole purpose of growth promotion can be discontinued.”

Low doses of antibiotics make farm animals grow slightly faster on less feed, presumably by easing minor infections that don't make animals overtly sick. But antibiotics spur some bacteria to develop resistance to the drugs, and those bugs, or the resistance genes they harbor, can make their way into meat and then the human gut. There, infections might withstand drugs akin to those used on the farm.

As antibiotic resistance spread in the last decade, the movement to phase out growth promoters gained traction. (Using antibiotics to treat sick animals or prevent infections from spreading has not been similarly criticized.) In 1997, WHO first recommended ending the practice. In 1998, Denmark—the largest pork exporter in the world—became the first country with a large livestock industry to phase it out, despite predictions from the animal-drug and livestock industries that the ban would lead to sicker animals, more contaminated meat, and economic damage to farmers.

To see how the “Danish experiment” has played out, scientists spent 4 years studying the prevalence of antibiotic-resistant microbes in farm animals, slaughterhouses, meat sold in groceries, and healthy people. They also tracked the impact of the ban on animal health and economic costs to farmers.

Overall, the ban accomplished its goals. Total antibiotic use in pigs and poultry was down 54% in 2001 from its peak in 1994. Drug-resistant strains in animals and meat fell dramatically; for example, 60% to 80% of Enterococcus bacteria samples from livestock resisted common antibiotics before the ban, whereas only 5% to 35% did so afterward. The data didn't show whether the phase-out reduced resistant bacteria in people, although other studies have revealed a link (Science, 4 January 2002, p. 27).

The ban had a slight downside for farmers, particularly pork producers. Farmers used about a third more antibiotics to treat sick animals after the ban. They had to treat just-weaned piglets for diarrhea about twice as often, piglets put on weight 2.6% more slowly, and slightly more died. Pig farmers' costs increased 1%, but overall profits from pork production continued to increase in Denmark.

Together, the data show that “there is no longer any need [for growth promoters]—if there ever was,” says microbiologist Stuart Levy of Tufts University School of Medicine in Boston, president of the Alliance for the Prudent Use of Antibiotics, a nonprofit group that includes researchers, doctors, and veterinarians.

Industry representatives beg to differ. The Danish ban hurt farmers without doing anything for public health, says Richard Carnevale of the Animal Health Institute in Washington, D.C., which represents manufacturers of animal antibiotics. He also faults the report for minimizing “significant economic costs to producers,” such as installing equipment to keep bacteria out of the barns and hiring extra staff.

European officials have already moved to end the use of growth-promoting antibiotics. By 1999, the E.U. had banned five drugs that are identical or closely related to human medicines. Last month, the E.U. finalized rules for its member countries to end the practice by 2006.

No such ban is on the table in the United States, where the Food and Drug Administration (FDA) rules on antibiotics on a case-by-case, use-by-use basis, says Stephen Sundlof, director of the agency's Center for Veterinary Medicine. FDA is also reexamining animal drugs approved before 1999, when the agency began formally considering public health effects.

U.S. reformers, frustrated by FDA's pace, have turned to Congress. In July, four senators introduced a bill that would stop livestock producers from using penicillins, tetracyclines, and six other antibiotics on entire herds or flocks unless the manufacturer proves the use is safe for human health. A companion bill has been introduced in the House, and 300 groups, including the American Medical Association, have endorsed the legislation.

Reformers have also made inroads in the private sector. In June, McDonald's Corp., which buys 1.1 billion kilos of beef, pork, and poultry each year, announced that it would insist that its chicken suppliers cut antibiotic use and would provide incentives for its pork and beef suppliers to do the same. “That was fantastic coming from a high-profile company like McDonald's,” Levy says. “We're waiting to see if Wendy's and Burger King will follow suit.”

4. ASTRONOMY

Panel Says NASA Should Consider Extending Hubble's Life

1. Robert Irion

The 13-year-old Hubble Space Telescope (HST) may be worth the extra money it would take to keep it functioning beyond 2010. That's the view of a panel of influential scientists, who last week* urged NASA to prolong HST's life with two upgrades during the next 7 years—provided that the second upgrade passes muster in peer review. However, such care would require costly space shuttle flights, an uncertain proposition as NASA juggles other expensive missions and evaluates the future of the shuttle program.

NASA asked the six-member panel, chaired by astrophysicist John Bahcall of the Institute for Advanced Study in Princeton, New Jersey, to review options for the $1.5 billion telescope as it nears the end of its originally scheduled operating life (Science, 20 June, p. 1857). The agency currently plans one more servicing mission in 2005 or 2006, the fifth orbital upgrade since the telescope's launch in 1990. Without a sixth visit, HST likely would expire by 2010 and plunge to Earth some years later in a controlled descent. However, spectacular recent results from HST—made possible by new instruments installed in March 2002 (Science, 22 February 2002, p. 1450)—have convinced researchers that the venerable telescope has plenty more science to offer if properly maintained. “Hubble is at its highest point in terms of demand and scientific productivity, and I believe that will continue as long as we service it,” says Steven Beckwith, director of the Space Telescope Science Institute in Baltimore, Maryland, which operates the instrument for NASA Moreover, an extended HST mission would overlap with the planned James Webb Space Telescope (JWST), a 6-meter telescope scheduled for launch in 2011. JWST will be sensitive to faint traces of infrared light far beyond HST's optical and near-infrared vision. “Great benefits accrue when major facilities study the same objects at different wavelengths,” says astronomer Garth Illingworth of the University of California, Santa Cruz. JWST's launch might slip past 2011, Townes and his fellow panelists noted. The panel also cited the advantages of having an uninterrupted stream of crisp images from space for scientists to analyze. “Continuity is very important,” observes Nobel laureate Charles Townes, a physicist at the University of California, Berkeley. “If there's a gap, science would be delayed, and the experts might go off in different directions.” The report concluded that HST's continued scientific potential and its major role in public appreciation of astronomy may justify the cost of a sixth repair mission in 2010. Accordingly, the oft-strapped Office of Space Science should not bear the burden by itself, the authors added. NASA officials estimate spending$600 million or more on the HST servicing team and instruments, plus $300 million to$600 million for the shuttle mission itself. That price tag prompted the panel to propose a novel approach to deciding whether to proceed: Force proposals for an extended Hubble flight to compete against other pitches for new astrophysics missions.

That approach has merit, says Anne Kinney, director of NASA's physics and astronomy division, although she's not sure how it would be implemented. “We try to compete like versus like, but we don't have anything else like [a sixth servicing mission],” she notes. Still, Kinney praised the panel for tackling less desirable options as well, including the specter of no future repair missions. “We need to be flexible,” Kinney says, pending NASA's review of the forthcoming report from the board investigating the Columbia shuttle tragedy.

Still, astronomers believe that the panel has brightened the outlook for a third decade of science for Hubble. “This was a well-chosen, independent group of well-respected people,” says Illingworth. “NASA has to take this report very seriously.”

5. WORKFORCE DIVERSITY

NSF, Academics Told to Act As If They Mean It

1. Jeffrey Mervis

What will it take to produce a more diverse U.S. scientific workforce? A lot of academic carrots and a few sticks, said participants in a daylong workshop sponsored by the National Science Foundation (NSF) last week. But no combination of incentives and penalties will succeed, they warned, unless universities take the problem more seriously and graduates can find more good jobs.

“We in academia are the problem,” declared Shirley Tilghman, a molecular biologist and president of Princeton University. “We have designed the career path in a way that discriminates against women and minorities.” She and others ticked off a long list of disincentives, including a lengthy apprenticeship, low pay, family-unfriendly hours, and cutthroat competition for grants.

The workshop supplemented a new report on national workforce policies from NSF's oversight body, the National Science Board, that laments what it calls an inadequate supply of domestic scientific talent (Science, 30 May, p. 1353). “We wanted to put greater emphasis on increasing the number of women and underrepresented minorities in science,” explains Warren Washington, board chair. “We hope that this workshop will lead to a statement by the board on how to achieve that goal.” Although the problem affects all of society, he adds, any new NSF policies will focus on academia, “because that's where we have the greatest leverage.”

An overflowing roomful of university presidents, government and professional society executives, and longtime advocates of greater diversity told the board that words are good but deeds are better. “If I ruled the world,” said Clifton Poodry, who oversees minority programs at the National Institutes of Health, “I'd give a lump [sum] of money to every university that is above average in sending minorities to graduate school. It would be a pat on the back, a message to keep up the good work.”

One major obstacle, says Tilghman, is that academic science can be a hostile climate for women and underrepresented minorities—defined as African Americans, Hispanics, and Native Americans. “Our great success story is the near-parity of women earning Ph.D.s. in the life sciences,” she noted. “But that's not the case for new faculty,” who are still predominantly male. “Is it that women aren't choosing academic careers? Or is it that academia is not choosing them? I think it's both.”

Tilghman said that university administrators can show they're serious about diversity by creating university-wide search committees with distinguished faculty members to seek out minority candidates—and by rewarding departments with additional resources. But others said that's not enough. “One [minority] faculty member isn't going to do it,” said George Langford, chair of the science board's education committee and a biology professor at Dartmouth College in Hanover, New Hampshire. Langford, an African American, noted that Dartmouth has not hired another minority in the natural sciences since he was recruited in 1991. And board member Pam Ferguson, a mathematician at Grinnell College in Iowa, said that one group of retiring white male professors in her department “saw no reason not to clone themselves” in choosing their successors—and would have succeeded had the college president not insisted that the process be opened up.

NSF has recently started to tackle the problem in an indirect way: It's cracking down on grant applicants whose proposals do not adequately describe the larger societal impact of their research, including steps to broaden the scientific pool by reaching out to underrepresented groups. But the effect so far has been minimal: Only 245 of some 30,000 proposals this year have been rejected because they fail to address the research's societal impact, says Nathaniel Pitts, a senior NSF administrator. Moreover, several participants scolded the foundation for not knowing if grantees really follow through on their promises, especially on big projects such as NSF's Science and Technology Centers (STCs).

“The STCs have done great things,” said Rice University mathematician and former board member Richard Tapia, who directed community outreach for an STC on parallel computing. “But they haven't changed the culture of the university.” Keith Jackson, president of the National Society of Black Physicists, was even blunter. “Is an STC going to get its funding cut off if it doesn't reach its goals for serving underrepresented minorities?” he asked. “That's what really matters.”

Jackson also took a swipe at the assumption in the science board's recent report that the demand for scientific talent is outpacing supply. “Our 600 African-American physicists have degrees from some of the finest universities in the country,” he said. “But there aren't jobs for them” after they graduate.

For a minority scientist on the bottom rung of the academic ladder, the climb can seem endless. “I know what I need to do to get tenure: publish copiously and bring in gobs of grant money,” said Emilio Bruna, a population biologist with 1 year under his belt at the University of Florida in Gainesville. “And being here to talk about diversity isn't going to help.” But being resourceful might. Bruna ended his remarks with a plea that transcends race and gender: “Give me a grant. Please.”

6. WORKFORCE DIVERSITY

1. Jeffrey Mervis

University programs that do a good job of training minorities and women scientists share nine essential traits, according to an upcoming report from a consortium of government officials, industrial leaders, and educators.

Next month, BEST (for Building Engineering and Science Talent) will send Congress its analysis of “what works” in elementary and secondary schools, higher education, and the workforce. But last week Shirley Ann Jackson, president of Rensselaer Polytechnic Institute in Troy, New York, summarized the report's academic segment for a National Science Foundation (NSF) workshop to promote diversity (see related story), adding an example of each trait:

• Institutional leadership (the Meyerhoff Program at the University of Maryland, Baltimore County)

• Targeted recruitment (the National Consortium for Graduate Degrees for Minorities in Engineering and Science)

• An engaged faculty (Preparing Future Faculty project)

• Personal attention (the Stevens Institute of Technology's Lore-El program)

• Peer and intergenerational participation (NSF's VIGRE and LSAMP programs)

• Comprehensive financial aid (institutional funding)

• Extended research experiences (the University of North Carolina's Partnership for Minority Advancement in the Biomolecular Sciences)

• Bridge programs to the next level (UCLA's Center for Excellence in Engineering and Diversity)

• Continuous evaluation (the Gateway Coalition based at Drexel University)

7. INFECTIOUS DISEASES

1. Dennis Normile,
2. Ding Yimin*
1. Ding Yimin writes for China Features in Beijing.

TOKYO—The masked palm civet, the obscure south Asian animal suspected of helping spread the virus that causes severe acute respiratory syndrome (SARS), is returning to Chinese markets. The lifting of a 4-month ban on it and 53 other species of wild animals delighted gourmets in Guangdong Province, where the local cuisine relies on a variety of wild animals. But it came as a surprise to the World Health Organization, which has a team of experts working with its Chinese counterparts investigating a possible animal reservoir of SARS.

The joint Chinese-WHO team “is trying to get hard evidence so one could do a risk analysis of which animals could harbor the virus” and transfer it to humans, says Alan Schnur, WHO's team leader for communicable disease control in Beijing. But any decision on setting or lifting marketing bans “is an internal matter [for China],” he says, “and WHO would not normally be involved in such a decision.” Civets have been identified as a potential link in the chain of infection.

China's State Administration of Forestry and the State Administration for Industry and Commerce banned sales of 54 species of wild animals in late April after researchers at Hong Kong University reported finding evidence of the SARS virus in several civets collected from a market in Guangdong Province, which borders Hong Kong. Although the researchers never advocated this action, officials decided to order a sales ban at the height of the SARS crisis as a precautionary measure. It dealt a devastating economic blow to thousands of families who raise the wild animals, and the families lobbied hard to have the ban reconsidered.

In early June, the forestry administration invited experts representing the Ministry of Science and Technology, the Ministry of Health, the National SARS Control and Prevention Command, and academic researchers to a symposium to assess the risks of allowing wild animals back into the markets. Guo Zhiwei, an official at the Ministry of Science and Technology, says the government “did not find any evidence of a connection to SARS among those [54] species.”

No other scientists provided any evidence of a link, either, says the forestry official, so the government lifted the national ban on farm-raised animals but retained a prohibition on sales of animals caught in the wild. Each province controls local commerce, and Guangdong has already declared that animal ranchers can resume sales of civets once quarantine examiners certify that their stocks are free of SARS and other infectious agents. The University of Hong Kong team that first made the link to civets emphasized in an earlier interview that it never claimed civets were the presumed animal reservoir, but that its findings suggested there is an animal link (Science, 18 July, p. 297). WHO's Schnur says it is understandable that the government would not want to base a policy decision on the preliminary findings of the University of Hong Kong and other teams. Resolving the question of an animal reservoir “is very much a work in progress,” he says.

8. INDIA

Mission Approved to Orbit Moon by 2008

1. Pallava Bagla

NEW DELHI—India has decided to send a spacecraft to orbit the moon and create a high-resolution topographical map of the lunar surface. With this mission, its first outside Earth orbit, India will join a decade of lunar exploration that is due to begin next month with the launch of Europe's SMART-1 probe (see p. 1031 and Science, 2 May, p. 724).

“I am happy to announce that by 2008 India will send a mission to the moon,” Prime Minister Atal Behari Vajpayee said last week in his annual Independence Day anniversary address. And its name, Chandrayan Pratham (the first journey to the moon), signals the country's intention to launch additional missions.

The $100 million mission, proposed 4 years ago by the Indian Space Research Organisation (ISRO), will place a 525-kilogram spacecraft in polar orbit 100 kilometers above the moon; it will carry instruments that work at wavelengths in the visible, near infrared, and low-and high-energy x-ray regions to explore the moon's composition. In addition to carrying payloads from several Indian labs, the spacecraft will include equipment designed by the Canadian and European space agencies. Although most scientists applaud the government's endorsement of the mission, some worry that the country isn't doing enough to ensure an adequate supply of astronomers, astrophysicists, and other scientists to take full advantage of such an ambitious space effort. “You need young and dynamic researchers to sustain a program as farsighted as this,” says Jayant Vishnu Narlikar, a theoretical astrophysicist and former director of the Inter-University Centre for Astronomy and Astrophysics in Pune. “For that to happen, a dialogue has to be initiated immediately with the universities.” Although ISRO has an annual budget of$600 million, it spends only about $2.5 million a year on extramural research. But V. Sundararamaiah, ISRO's scientific secretary, says the agency hopes to create a dozen or so junior research fellowships at select institutes and universities to serve the needs of its planetary exploration program. 9. ECOLOGY Lean Winters Hinder Birds' Summertime Breeding Efforts 1. Jocelyn Kaiser SAVANNAH, GEORGIA—Migratory songbirds have become scarcer in recent decades, in part because their tropical wintering grounds are being degraded and many birds don't survive the winter. A new study uses a chemical marker in birds' blood to suggest that this habitat loss has a ripple effect in surviving birds that extends well into the breeding season, when the birds may be a continent away. The American redstart, a warbler, spends winters in the Caribbean, Central America, and northern South America and summers in temperate forests in the United States and Canada. There the birds mate and produce up to five chicks. Redstarts that winter in poor habitat produce fewer chicks and their chicks fledge later, a team led by graduate student Ryan Norris of Queens University in Kingston, Ontario, reported here earlier this month at the annual meeting of the Ecological Society of America. The finding builds on earlier work linking winter habitat with a bird's health when it arrives to breed. But “nobody has demonstrated” a direct link between winter habitat and breeding success until now, says avian ecologist Susan Hannon of the University of Alberta in Edmonton. “This adds another potential problem [birds] have to deal with,” along with threats such as fragmentation of northern forests, which attracts cowbirds that replace redstart eggs with their own. A few years ago, Peter Marra of the Smithsonian Environmental Research Center in Edgewater, Maryland, and co-workers developed a chemical technique that can identify the kind of habitat in which a bird overwintered. Because plants use different pathways for photosynthesis, those that grow in richer, wetter tropical habitats—such as mangroves and wet lowland forests—contain less of the carbon-13 isotope than do plants in drier areas, such as scrub. Insects eat the plants, birds eat the insects, and the habitat leaves a carbon-13 signature in the birds' blood. Marra's team used this marker to show that redstarts arrive up north sooner and in better physical condition when they have spent the winter in a richer habitat (Science, 4 December 1998, p. 1884). In the new study, Norris, his adviser Marra, and co-workers show that better winter habitat translates into better breeding success. Norris spent two summers monitoring about 90 male and female redstarts nesting north of Lake Ontario. He found a striking correlation between carbon-13 levels in the birds' blood and breeding success: Males that had better winter diets not only arrived earlier at the breeding grounds but also sired slightly more young. Effects were even stronger for females: Those that arrived from better habitats produced up to two more chicks and fledged them up to a month earlier than did females that wintered in sparser grounds. “Negative effects in one season can be negative again in another season,” Norris says. He says this “carryover” effect underscores the fact that conserving migratory birds will require saving wet tropical forests and mangroves, which are rapidly being lost to logging and development. 10. OCEAN EXPLORATION Deep-Sea Mountaineering 1. David Malakoff Marine scientists are traversing the oceans to survey and explore the diverse fish, coral, and other organisms living on thousands of submerged peaks ABOARD THE ATLANTIS IN THE NORTH ATLANTIC—The dazzling red squid drifted into the submarine's spotlight like some psychedelic phantom, its supple tentacles pulsing against the cold, black sea. For a mesmerizing moment, the cephalopod and three human explorers hung eye to eye more than a kilometer down. Then the squid jetted away, out of sight—but not out of mind. “How often do you get to visit a resident of the deep sea in its own home?” exclaimed an exuberant Peter Auster just hours after the encounter, which occurred during an expedition last month to several drowned volcanoes off the coast of New England. “It was unforgettable.” Auster, a fish specialist at the University of Connecticut's National Undersea Research Center in Groton, is part of a growing cadre of undersea mountaineers. Around the globe, ocean scientists are launching an unprecedented wave of expeditions to survey and explore underwater mountains that tower up to 4 kilometers above the ocean floor. The rush to scale these peaks, called seamounts, is prompted in part by growing evidence that they are home to remarkable numbers of new and unusual organisms. Biologists also want to document seamount ecosystems before they are damaged by fishing trawlers or other human activity. Conservationists, meanwhile, are pushing governments and the United Nations to protect these underwater summits. Researchers have visited just a few hundred of the world's estimated 30,000 seamounts. Fewer than 1000 even have names, and only a handful have been intensively studied. But that's changing fast. The United States, Europe, New Zealand, and other nations have begun funding cruises—such as last month's U.S. expedition to three peaks in the New England chain—to close that knowledge gap. This week researchers are meeting in Oregon to discuss scientific priorities and how to pool their findings. Such planning comes none too soon, says Les Watling of the University of Maine's Darling Marine Center in Walpole, chief scientist on the recent expedition. “Seamount biology is in its infancy, but it needs to grow fast if we're going to have an impact,” he says. Chain attraction Drowned mountains are the stuff of legend. And although century-old soundings of the ocean floor provided the first data on the extent of these submerged mountain chains, the term “seamount” wasn't introduced until 1936, when the U.S. government officially used it to name the massive Davidson seamount off Southern California. Today, hydrographers define a seamount as an independent feature that rises at least 1000 meters above the sea floor. The Pacific Ocean is home to about two-thirds of them, with the rest scattered across the other oceans. The peaks are typically extinct volcanoes and, until recently, the bulk of seamount research focused on their geologic origins. Indeed, geoscientists still debate the mechanisms that allow magma to punch through the sea floor. The search for new fishing grounds in the late 1960s prompted biologists to join the fray. The Soviet Union and Japan, in particular, dredged dozens of previously unknown species from subsurface peaks to give marine biologists their first taste of seamount biodiversity. But trawlers often followed in their wake, decimating fish populations. It took Asian ship captains less than a decade in the 1970s, for instance, to sweep armorheads, a type of deep-water fish, from Hawaiian seamounts. More recently, alarm over such practices grew after researchers in New Zealand and Australia began to study the impact of the lucrative fishery for orange roughy, a long-lived deep-sea fish that schools above seamounts. Particularly influential were papers by marine scientist J. Anthony Koslow of Australia's Commonwealth Scientific & Industrial Research Organisation in Hobart and colleagues elsewhere. They reported that trawling nets, besides eliminating up to 90% of orange roughy populations, had crushed gardens of long-lived corals and stripped slopes nearly bare. Some trawled seamounts, for instance, show 95% bare rock compared with just 10% for unfished summits. Similarly, fished seamounts supported half the biomass and far fewer species than those that hadn't been trawled. Such findings “opened a lot of eyes to what was happening out of sight,” says Karen Stocks, a University of California, San Diego (UCSD), researcher who has helped develop Seamounts Online, an Internet database. Koslow and other researchers, meanwhile, were confirming long-held hunches that some seamounts support unusually productive ecosystems. In the late 1980s, Raymond Wilson of California State University, Long Beach, and Ronald Kaufmann of the University of San Diego concluded from the few existing surveys of seamount organisms that up to 15% were “endemic,” or limited to just one or a few adjacent peaks. More recent studies have found even higher rates, with French biologist Bertrand Richer de Forges of the Institute of Research for Development in Noumea, New Caledonia, reporting that one-third of the species found on some seamounts off New Caledonia are endemic. Koslow has found similar rates off Tasmania, and Nikolay Parin of Russia's Shirshov Institute of Oceanology in Moscow estimates that up to half of the fish and invertebrates on the Nasca and Sala-y-Gomez seamounts off Chile are unique. “Almost every time you visit a new seamount, you find something unique,” says marine scientist George Boehlert of Oregon State University's Hatfield Marine Science Center in Newport. Along with such data came clues that some seamounts also serve as steppingstones for species to expand their ranges. Genetic data suggest that some species have jumped from seamount to seamount over thousands or millions of years. Seamounts may also help individual animals—particularly large predatory fish—navigate long distances. Some researchers believe that migrating hammerhead sharks, for instance, use seamounts as rest stops and even rely on magnetic fields surrounding the features as guideposts. These findings led many scientists and conservationists to a worrying conclusion: Human activities could soon wreak havoc on ecosystems that give birth to new marine species and provide refuge to vulnerable organisms. That fear has spurred a lobbying effort to increase seamount research and conservation. Into the deep Last month's 10-day cruise aboard Woods Hole Oceanographic Institution's (WHOI's) Atlantis is one manifestation of that increased concern. The ship, with a dozen scientists and a roughly equal number of reporters, science teachers, and government officials on board, visited three of the roughly 35 major seamounts in the New England chain, which stretches 1500 kilometers southeast from Boston into the Atlantic. Scientists had already paid at least one visit to two of the peaks, Bear and Manning. But the Atlantis crew was the first to glimpse Kelvin, a 3400-meter-high plateau near the center of the chain whose peak sits 1600 meters below the surface. Demystifying such features is a priority for the National Oceanic and Atmospheric Administration's Office of Ocean Exploration, which funded the$550,000 cruise as part of a multiyear seamount research plan.

Like most oceanographic expeditions, this one had a round-the-clock agenda. Tasks ranged from using Atlantis's sophisticated multibeam sonar to create detailed three-dimensional sea-floor maps to unsuccessfully attempting to collect live deep-sea shrimp for a researcher hoping to raise them in tanks (see sidebar, p. 1036). Researchers also collected fossil corals for a climate researcher.

Two goals dominated the cruise, however. One was to survey deep-sea corals and collect specimens for genetic and reproduction studies. The other was to observe how fish use the rugged seascape. The coral specialists were headed by Watling and Scott France of the College of Charleston, South Carolina, and the fish experts were led by Auster, Jon Moore of Florida Atlantic University in Jupiter, and the University of Connecticut's Ivar Babb. They also had the luxury of using WHOI's deep-diving submarine, Alvin, to get a firsthand look at their subjects rather than relying on nets or robotic cameras.

“We've never observed a lot of these beasts alive in their own habitats. They're usually dumped dead on some deck,” Auster noted before the first of seven dives to 1500 meters or more. Moore and Watling, meanwhile, were looking forward to seeing how seamount organisms are distributed across the slopes. “It's hard to tell what lives with what when they are jumbled together in a net,” says Moore. “Alvin should help us see the big picture.”

By the end of the cruise, such hopes had been fully realized. Each day, the stubby, three-person submarine returned to the surface laden with brightly colored corals, some with delicate brittle stars entwined in their branches. Sometimes, the “bioboxes” —Plexiglas containers designed to keep specimens cool and alive—held oddities, such as the large leathery egg of a skate, or a small orange shrimp with one huge claw. There were also dozens of hours of videotape, capturing everything from the close encounter with the meter-long cherry-red squid and wriggling fish, to tangerine-colored, 2-meter-long whip corals corkscrewing in deep currents. Still images captured seamount vistas that resembled alpine meadows, with bare stretches of basalt separating patches of gangly corals and vase-shaped sponges.

Each night, the researchers worked late preparing and packing the specimens for museums or the lab. All of the finds will be put to use. France and graduate student Mercer Brugler, for instance, are studying the genetic diversity and distribution of black corals, a poorly understood group. Other samples may help France understand why some deep-water octocorals—a group that includes the feathery sea whips and sea pens—have so far shown almost no variation in their mitochondrial DNA, which often holds markers used to divide other organisms into distinct populations.

Watling and graduate student Anne Simpson, meanwhile, hope to use anatomical studies to shed light on how octocorals, some of which may live 100 years or more, feed and reproduce. And Moore, an expert on the biota of the New England seamounts, will be combing through his data to refine existing theories on where the seamount's deep-sea fish originated. In recent work, he and colleagues have concluded that about 85% of the roughly 600 known species came from south of the chain, carried by the Gulf Stream or other north-trending currents. Most of the rest originated in the subarctic, perhaps carried south by deep currents.

Most intriguing, however, are a few species previously known only from the eastern Atlantic off Europe. Moore says it's a mystery how these fish moved across prevailing currents and vaulted the towering Mid-Atlantic Ridge—a process akin to crossing a busy highway. One of these migrants—a diamond-shaped fish in the Oreosomatidae family—was spotted on Bear seamount and apparently has never been seen so far west. Such finds demonstrate “how little we know about deep-sea biogeography,” he says.

International interest

Atlantis wasn't the only vessel visiting seamounts this year. In the Pacific, a joint New Zealand-Australian expedition this spring explored more than a dozen peaks near those nations. For New Zealand, the cruise was the latest chapter in a long-term research initiative begun in 1999 by the National Institute of Water and Atmospheric Research that will run through at least 2006.

For Australian researchers, this year's joint cruise advanced efforts begun in 1997 in waters off Tasmania. Findings from that expedition helped build support for the 370-square-kilometer Tasmanian Seamounts Marine Reserve, created in 1999 and encompassing more than a dozen summits. Australian researchers have also benefited from the work of French scientists, who have been exploring nearby seamounts since the mid-1980s.

European scientists also went seamounting in their own waters this year as part of OASIS, a 5-year European Commission project aimed at sizing up several summits in the northeast Atlantic. A 19-nation project focused on the Mid-Atlantic Ridge, dubbed MAR-ECO, also is under way.

All this activity has prompted discussions on how best to share the expected flood of data and how to set research priorities. At this week's seamount summit at Oregon State University, sponsored by the Census of Marine Life research consortium, scientists will attempt to compile a set of questions that might guide them for the next 5 to 7 years.

Aboard Atlantis, such overarching questions abounded. Watling, for instance, pondered whether factors—such as water depth, current patterns, or distance from the coast—could explain the location and diversity of the coral communities he saw. The researchers also discussed whether the seamount organisms were actually reproducing on the summits or drifting in from other areas. If the populations were sustaining themselves, they wondered whether circular seamount currents known as Taylor cones were preventing eggs and larvae from drifting into the nearby abyss.

To help answer such questions, UCSD's Stocks has been urging researchers to share data through electronic databases such as Seamounts Online. To draw robust conclusions, she says, “you need to bring together data from a lot of seamounts, not just one or two.” That goal got a boost recently when Russia agreed, with U.S. help, to contribute 2 decades of cruise results to the site. Another site, meanwhile, is collecting seamount maps, starting in the Pacific (see NetWatch, p. 1025).

Other scientists hope the Oregon meeting will spur coordinated research. For instance, marine biologist John Dower of the University of Victoria, Canada, would like to see researchers conduct studies that compare seamounts with similar physical characteristics but different biological productivity. Such comparisons could help reveal why some seamounts have higher rates of endemism or richer fish communities than others.

Dower's own studies in the Pacific off Canada may have already answered one puzzle: How do isolated seamounts support fish populations that appear to be genetically identical to distant inshore stocks? Studies suggest that in the northeast Pacific, the answer is that giant eddies periodically swirl off the coast, carrying larval fish out to the peaks. Such findings have implications for conservation, he adds, because “you couldn't just protect the seamount; you'd need to protect the source [of the fish] too.”

Similarly, “policymakers will want to know if bottom trawling is going to be a problem on every seamount or just some,” says conservationist Matthew Gianni, an Amsterdam-based adviser to the International Union for the Conservation of Nature, who has been urging nations to impose a moratorium on seamount trawling. His goal is a global agreement, and he hopes that a U.N. ban on high-seas drift netting a decade ago will serve as a precedent. Several major conservation groups, including Conservation International and the World Wide Fund for Nature, have taken up the cause.

Researchers also hope to raise the profile of seamounts with Web sites that let the public experience scientific cruises vicariously through pictures and daily updates. “Seamounts are big, photogenic features,” says Gianni. “We just need to bring them out of hiding.” Such exposure, researchers hope, will help turn the current molehill of seamount data into a mountain of useful knowledge.

11. OCEAN EXPLORATION

Bringing 'em Back Alive

1. David Malakoff

Most seamount researchers are content to return from their expeditions with heaps of preserved specimens or hours of video tape. A few, however, seek more vibrant souvenirs: deep-sea creatures that can be kept alive in their laboratories for months or even years.

One might think that the rapid change in pressure would doom these denizens of the deep. But it turns out that fluctuations in water temperature are more of a threat, says marine biologist James Childress of the University of California, Santa Barbara, who has been bringing deep-water mussels, crabs, and other organisms back alive for 3 decades. “Most deep-sea animals have very narrow temperature tolerances,” he says, and thrive in waters a kilometer or more down that typically hover at just 2° to 4° Celsius.

To avoid deadly thermal shock, the captured creatures are placed in devices—from insulated net ends to special collecting boxes (dubbed “worm coffins” by some)—that keep them surrounded by cold water on their way to a surface that might be 20 degrees or more warmer. Once transferred to chilled aquaria, some creatures have lived for more than 2 years.

To be sure, some deep-sea animals can't take the sudden drop in pressure. Atmospheric pressure at the surface is about 100 times less than at 1000 meters down. The difference can cause deep-sea fish with swim bladders to literally blow up as their body gases expand. And scientists have found that cell membranes and enzymes of creatures collected below 2000 meters often don't function properly at surface pressures. Interestingly, however, the reverse isn't always true. Some coastal mussels seem to do just fine when dropped into deep waters, for instance, although researchers aren't sure how they adapt to the crushing pressure.

12. PLANETARY SCIENCE

Eons of a Cold, Dry, Dusty Mars

1. Richard A. Kerr

Evidence is mounting that, although water has at times flowed across the martian surface, it has not lingered as a liquid long enough to alter the planet much chemically

PASADENA, CALIFORNIA—The trio of missions that will touch down on Mars this December and January are “following the water” as they search for signs of life on another planet. Each targets an ancient site where water may once have stood, perhaps during an oft-hypothesized warm and wet era on Mars. But the spectral data pouring back the last few years from instruments orbiting Mars are starting to build a case for a volcanic planet only gently touched by the chemical alteration that water has wreaked on Earth for billions of years. In the emerging view, a perpetually frigid climate has kept Mars's modest store of water locked up as ice most of the time. The surface may never have been particularly hospitable to life, at least not for long. If so, the search for signs of life will have to focus on the martian interior and its rare surface exposures.

Planetary scientists presented new evidence for a long-cold, long-dry Mars here last month,* and a report in this issue of Science (p. 1084) provides additional support for that notion. Researchers are finding minerals long exposed on the surface that would have been quickly destroyed by liquid water. And the products of extensive weathering expected under a warm, wet climate, such as clays, are showing up in unexpectedly tiny quantities, if at all. “That argues it's a dominantly dry planet,” says planetary geologist and spectroscopist Philip Christensen of Arizona State University (ASU) in Tempe, and it has almost always been that way. Water does seem to have flowed on the surface briefly, early in martian history (Science, 6 June, p. 1496). And it has gushed to the surface in the planet's midlife and seemingly trickled down gullies in the geologically recent past (Science, 28 February, p. 1294)—perhaps as a tiny bit of present-day “ice-ball Mars” melted at higher latitudes (Science, 11 April, p. 234). But “it hasn't been warm and humid and tropical,” says Christensen, at least not long enough to leave its mark on the rocks.

Christensen has been pushing the strongest version of the cold, dry Mars hypothesis. He has argued that, so far, remote sensing has shown the surface to be unaltered basalt, the volcanic rock of Earth's ocean crust, and unaltered andesite, a more silica-rich volcanic rock. There are, he concedes, a few small areas that look like once-buried aquifers. In support of a Mars unaltered by water, he points to olivine, the greenish mineral found in basalt and recently recognized on the surface. With a modicum of moisture at even low temperatures, olivine should crumble away in just a few thousand years, according to laboratory work reported at the meeting by Julie Stopar of the University of Hawaii (UH), Manoa, and her colleagues.

Yet this year, using data from the Thermal Emission Imaging System (THEMIS) orbiting on the Mars Odyssey spacecraft since 2001, Christensen, Roger Clark of the U.S. Geological Survey in Denver, Colorado, Victoria Hamilton of UH, and others have reported the detection of olivine-rich basalt at several places on Mars. For example, thanks to the high spatial resolution of THEMIS, Christensen can point to a 100-meter-thick layer of olivine-rich basalt exposed near the bottom of the gorge of Ganges Chasma. The olivine is “really old,” says Christensen. It has been at or near the surface for hundreds of millions of years, he guesses, to judge by the age of the gorge that has exposed it. Its survival argues “for a pretty dry Mars,” he says.

At the same time, spectroscopists are failing to find the clays that result when basalt weathers in damp warmth. At the meeting, Steven Ruff of ASU reported that spectral data returned by the Thermal Emission Spectrometer (TES) operating onboard the orbiting Mars Global Surveyor since 1997 cannot rule out the presence of clays on the surface. But his latest analysis pushes the potential abundance down to a few percent at most. “We have extremely good spectra of the dust,” says Clark. “We haven't detected the weathering products, particularly the clays.”

Researchers are reporting subtle signs of other weathering products, but their presence is consistent with the limited weathering expected on a cold, mostly dry Mars. In the latest development, Joshua Bandfield and Timothy Glotch of ASU and Christensen identified the mineral carbonate in TES spectra, as they reported at the meeting and describe in this issue. “The quantity of TES observations has allowed us to develop techniques to correct for all the [interfering effects of] dust, water vapor, and carbon dioxide in the atmosphere to give us a high-quality spectrum of the dusty surface,” says Bandfield. The wiggles and squiggles of that spectrum are “consistent with a small amount of carbonate,” he says, about 2% to 3%.

On Earth, water and atmospheric carbon dioxide combine to form carbonic acid that eats away at rocks and flushes their remains into lakes and oceans. There, dissolved carbonates will eventually precipitate to form solid carbonate deposits like the White Cliffs of Dover, with or without the help of animals that form carbonate skeletons. But geochemists imagined that wet weathering would produce something like 20% carbonate in martian dust, not 2%. That's no more than might form over the eons by the dust's reaction with today's vanishingly low martian humidity, notes Christensen, or the higher humidities possible during the ebb and flow of martian ice ages. And he adds that the TES detection of scarce carbonate in dust only reinforces the instrument's failure to detect any exposed carbonate rock deposits on Mars larger than an area of a few tens of square kilometers. TES isn't somehow failing to see sizable carbonate deposits, he says, they just aren't there.

At the meeting, planetary spectroscopist Richard Morris of NASA's Johnson Space Center in Houston reported another sign that Mars has weathered only moderately. Because TES operates at long, thermal-infrared wavelengths, it samples beneath the rock's surface, up to 10 micrometers in. At less penetrative visible and near-infrared wavelengths, where the Hubble Space Telescope (HST) operates, spectra reflect the composition of just the outer micrometer or two.

In visible and near-infrared images taken by HST, Morris noted, dark areas of Mars have clear signs of a weathered rind. Because TES doesn't see it, he says, it must be only a micrometer or two thick. Spectrally, it resembles the clay-free material called palagonite that forms on moderately—but not severely—weathered basalt on Earth. On Mars, the yellow-brown palagonite could form under cold, humid conditions over the eons and flake off to add to the planet's soil and ubiquitous dust, says Morris, giving the planet its tint. The visually bright areas on Mars, including the northern lowlands, do spectrally resemble such palagonitic debris, he says, although he can't rule out Christensen's unaltered andesite there.

Other signs of weathering under relatively dry conditions were reported at the meeting. Spectroscopist Scott Murchie of the Johns Hopkins University Applied Physics Laboratory in Laurel, Maryland, reanalyzed spectra returned by the Mars Pathfinder lander in 1997 and found that some rocks at the landing site have a thin coating of “desert varnish” of the sort that forms in the desolate Dry Valleys of Antarctica. Even monomolecular layers of dew over a few millennia could have cemented a thin, iron-rich dust layer onto the rock to give a similar spectral signature, says Murchie. But only the rocks that have been exposed the longest, as suggested by their positions beneath rocks that arrived later as impact ejecta, have the varnish. That would put the more humid era required for varnish formation back a billion years or more, says Murchie.

This emerging spectral evidence presents a problem. “We have a real contradiction between the mineralogy and the geomorphology,” says planetary scientist Bruce Jakosky of the University of Colorado, Boulder. Geologists are seeing more and more evidence that water cut networks of valleys during earliest Mars history, and water seems to have cut small gullies throughout martian mid-latitudes in recent geologic times. “If water is that abundant,” asks Jakosky, “where are the weathering products?”

A nearly perpetually cold Mars is one possible answer. Cold could lock up water as ice almost all the time except when a particularly extreme swing of the planet's axis brings extra solar heating to polar regions. Then the water could move to lower latitudes as water vapor and snow. Subsequent melting of that snow at the height of rare summer days could then shape the landscape as seen today, before an easing of the tilt returned Mars to a complete deep freeze.

At the meeting, spectroscopist Michael Wyatt of ASU reported an intriguing spectral pattern that could reflect such an ice-ball Mars. He showed how, in both the northern and southern hemispheres of Mars, the intensity of weathering as gauged by the strength of the palagonite-like spectral signature varies by latitude. Apparent weathering is greatest near the polar ice cap, where subsurface ice was recently found. It declines across mid-latitudes, where intermittent near-surface ice has recently been suggested, and is lowest in the martian tropics, where ice is least likely ever to have formed. That pattern fits an ice-ball Mars.

Even spectroscopists do not yet trust their squiggly lines to tell them all about Mars, however. Upcoming missions will take a closer look. Earlier this month NASA announced plans to land a soil and ice analyzer on the ice-laden northern plains of Mars in 2008 (Science, 8 August, p. 743). And three Mars landers—two NASA and one European—are now on their way to the planet. A U.S. rover is headed for what appears to have been a wet, buried hot spot. The other will investigate an ancient lakebed. Both carry a bevy of analytical instruments, including a “mini-TES” for a close-up look. The European Beagle 2 lander will come down on what may be an ancient sea floor. With luck, they'll have a firsthand chance to spy water's work on Mars.

• *Sixth International Conference on Mars, 20 to 25 July, California Institute of Technology.

13. GENETIC PRIVACY

Japan Guidelines Under Fire After Protest Halts Study

1. Dennis Normile

The Japanese Medical Association has raised concerns about government oversight after city workers distributed questionnaires for a health study

TOKYO—When cancer epidemiologist Kei Nakachi enlisted municipal employees to help him gather data to study the genetic and lifestyle components of cancer in a small Japanese town, he stressed the importance of preserving their neighbors' privacy. But last month his ethical oversight committee abruptly suspended the study, begun last May, after the Japan Medical Association (JMA) accused Nakachi and the Ministry of Education, his funding agency, of failing to ensure the confidentiality of medical and genetic information. The ministry's overall approach to handling genetic information “is in urgent need of reform,” the association said in a protest letter that pointedly mentioned several other large-scale human genetics studies now gearing up.

JMA's broadside sent a shiver through a research community still learning to follow ethical guidelines issued by the government 2 years ago. Ken Yamaguchi, president of the Shizuoka Cancer Center and head of the working group that drafted the guidelines, says that everyone anticipated occasional revisions. “But this is not something for JMA to address on its own,” he says.

JMA isn't backing down, however. Next week it will convene a committee to investigate the government's handling of genetic information in research. “JMA has an obligation to take an active role in protecting patient privacy,” says Hidetoshi Nishijima, an executive member of JMA's board of trustees.

All sides agree that Japan was slow compared with other industrial countries in adopting ethics guidelines for genetic research. The March 2001 rules are similar to standards elsewhere in detailing the responsibilities of institutional review committees, spelling out procedures for procuring informed consent, and requiring that, in addition to review boards, institutions “appoint a personal-information administrator to ensure protection” of individual privacy. Yamaguchi calls them “the strictest in the world” in considering the privacy of individual genetic data.

JMA disagrees. The guidelines “are a bit insufficient,” says Nishijima. They don't specify who owns genetic information, and they leave too much to the discretion of review committees, he says. Those shortcomings are exacerbated, he adds, by exempting academics from the provisions of a recent law intended to prevent the improper use of personal and financial information in computer databases. This attempt to preserve academic freedom is not an escape hatch, Yamaguchi counters, because academics can still be censured and lose funding if they break the rules. Medical professionals already face legal penalties for breach of confidentiality, he adds.

The municipal workers collecting data for Nakachi's survey fall outside both categories, however. The town of Kumanocho, outside Hiroshima, hired 130 people to reach the 14,000 residents over the age of 40 involved in the study. The workers distributed a written explanation of the study along with a lifestyle and medical history questionnaire to be collected later. Nakachi then planned to invite interested citizens to an explanatory meeting and obtain written consent from any volunteers before taking blood samples.

Although many researchers believe that Nakachi was wrong to deploy such a workforce, they are puzzled by JMA's criticism of other research projects. Tohru Masui, a developmental biologist who now studies biomedical research policy issues at the National Institute of Health Sciences in Tokyo, says he shares JMA's concerns about patient privacy but believes that its complaints are also “a way of winning a greater say for physicians in planning and reviewing such projects.”

In its 16 July letter to the Minister of Education, JMA specifically mentioned a ministry-sponsored Biobank that will contain genetic information on 300,000 individuals as well as the international Haploid Map project, which originated at the U.S. National Institutes of Health. Like similar biobank projects in other countries, Japan's hopes to link single-nucleotide polymorphisms to diseases and adverse drug reactions as a step toward tailoring medical treatment to genetic characteristics (Science, 8 November 2002, p. 1158). “The ministry's crudeness and highhandedness in the collection of materials and data for such medical research has become conspicuous,” the letter claims.

The charges are unfounded, says the Biobank's principal investigator, Yusuke Nakamura, a physician and geneticist at the University of Tokyo's Institute of Medical Science. The Biobank will ensure privacy through the use of separate databases, encoding, and firewalls. “It will be impossible to connect the genetic information to any individual,” Nakamura says. But he still worries that JMA's accusations may interfere with the recruitment of subjects later this year.

A compromise may be in the offing, however, thanks to the intervention of a prominent medical school. JMA president Eitaka Tsuboi, Nishijima, and several other key JMA officials are all graduates of Nippon Medical School, one of Japan's most prestigious schools for physicians. The school is cooperating with the Biobank plan in recruiting patients. Mitsuru Emi, a surgeon and professor at the school, says, “We are making every effort [to bring the two sides together] so that the misunderstanding between [the Biobank] project and JMA gets resolved.”

Nakamura is now awaiting a formal invitation to present details on the Biobank's privacy safeguards to JMA's new committee. The outcome is unclear. But the panel's debut marks the start of another round of scrutiny for researchers planning similar studies.

14. BIOINFORMATICS

Gene Counters Struggle to Get the Right Answer

1. Elizabeth Pennisi

Until researchers determine what constitutes a gene, they can't tally up how many humans have. In the meantime, gene-hunting programs are becoming more sophisticated

Researchers have been counting human genes for decades, but the numbers just don't add up. The best estimate soared to 100,000 a few years ago, dropped to about 30,000 when the human genome sequence was published, and recently sank as low as 20,000. To take full advantage of the sequencing of the human and other genomes, researchers say, they need a better accounting.

In more optimistic times—a mere 3 years ago—the genome-sequencing community started a betting pool called GeneSweep on what the number of human genes would turn out to be once the sequence was finished. This summer, researchers admitted they were nowhere near establishing a final count. They decided to end the suspense, saying that the books balance out at 24,500 protein-coding genes for now and naming a winner (Science, 6 June, p. 1484).

The problem, says David Kulp, a computer scientist at Affymetrix in Santa Clara, California, is that when it comes to defining a gene, “it's very difficult to say definitely what's right or wrong.” Molecular biologists are finding that some genes are shorter than anybody expected a gene to be. Sometimes it's hard to tell whether a piece of code is a single gene or two that overlap. And the community is not quite sure how to classify genes that code for multiple proteins or genelike sequences that code only for RNA.

This complexity has taxed bioinformaticists to the limits of their software-writing abilities. People in this decade-old field design computer programs to analyze DNA sequence data, which includes detecting genes. Their mathematics are increasingly sophisticated, with algorithms that take into account the geneticist's best knowledge of genes and proteins, as well as the molecular biologist's insights into how genes are hidden in DNA. Some of these computer buffs have even started doing their own experiments to characterize genes better.

They have a lot of work to do. Often they can tell that a stretch of DNA codes for an amino acid sequence, but the size, number, and exact distribution of protein-coding and noncoding regions in that gene remain elusive. Most worrisome to some is so-called dark matter, seemingly geneless regions in a genome that might contain hidden coding sequences.

Elusive prey

In the 1930s, George Beadle and Edward Tatum suggested that each gene codes for just one protein, an assumption that remained the conventional wisdom for decades. Now it's known to be oversimplified. One gene can yield multiple proteins or even be transcribed into RNA rather than a protein.

The protein-coding regions of human genes, called exons, take up only 2% of the DNA and can get lost in the other 98%. Genetic oddballs complicate gene counts as well. Some very simple genes consist of just one exon. They are so small that they are easily overlooked by both human and computer gene counters. In contrast, genes that no longer function because of some aberration in their DNA—so-called pseudogenes—artificially inflate gene numbers.

Among the 24,500 genes in the current assessment, “3000 could be pseudogenes,” points out Ewan Birney, one of the chief gene counters at the European Bioinformatics Institute in Cambridge, U.K. And he's not the only one who is stuck trying to decide which genes are real. “I believe all gene-prediction programs suffer from this,” says Michael Brent, a computer scientist at Washington University in St. Louis. “Everyone will do better [at their predictions] once we get the pseudogenes taken care of.”

Even worse, parts of the genome have proven completely impenetrable to the best gene-prediction programs. They include dark matter, regions named because they “are apparently devoid of genes,” says Roderic Guigó, a computational biologist at Pompeu Fabra University in Barcelona, Spain. Gene-prediction pros know nothing about this dark matter. Many worry that this void contains genes that researchers just can't see; dark matter genes “may have characteristics other than the ones we recognize,” Guigó points out.

One gene … two genes …

Any gene-prediction program worth its computer time must do a decent job of finding genes outside the dark matter zone. Typically they do this in one of two ways. The “ab initio” approach recognizes genes by detecting distinctive patterns in DNA sequences, such as those that characterize the beginnings and ends of genes. The other approach is comparative: It uncovers new genes based on their similarity to known proteins and genes. The two create a Goldilocks dilemma. Ab initio programs classify anything that looks vaguely like a gene as a gene, so their totals are too high. Comparative approaches don't recognize unfamiliar genes, so their estimates are too low. And nothing seems to estimate gene numbers just right. “But the programs perform a lot better than they used to,” says Gary Stormo, a computer scientist at Washington University in St. Louis.

Many people trace the field's beginnings to an ab initio program called Gene Modeler, produced in 1990 by Chris Fields and Cari Soderlund, who were then at New Mexico State University in Las Cruces, to find genes in the nematode Caenorhabditis elegans. Other software in existence at the time was much clunkier and took less direct approaches. For example, BLAST and FASTA translated DNA sequence into protein sequence that could be compared to existing protein data.

The field grew quickly. Other early predictors included Guigó, who adopted Gene Modeler's approach to build GeneID for finding human genes instead of worm ones. In 1991, one of Stormo's graduate students, Eric Snyder, wrote software called GeneParser that incorporated a technique called dynamic programming to separate exons from introns, gene regions that don't code for proteins. It worked more efficiently than other approaches by allowing the computer to consider just subsets of the data as it evaluated sequences.

Snyder, now at Pennington Biomedical Research Center in Baton Rouge, Louisiana, let the project lapse. “If I were to do it over, I would have kept working on GeneParser and gene prediction,” he says. But at the time, granting agencies didn't think the problem of counting genes was particularly important, and he was not alone in leaving his software behind for other projects.

A few have been lucky enough to be in the right place for keeping up their gene prediction work. Steven Salzberg and his crew at The Institute for Genomic Research (TIGR) in Rockville, Maryland, have been improving their programs for finding human genes since their first one, an ab initio approach, came out in 1994. They have recently come up with several new programs, one of which incorporates more background information to generate predictions, such as clearer rules about sizes of exons and introns. Another program works with two whole genomes at once, computationally laying one on top of the other for comparison.

Many researchers are taking this latter approach because similar species tend to have genes with very similar sequences. Protein-coding regions are likely to match and thus stick out among the unmatched nonsense sequence surrounding them. Not only genes match, says Eric Green, a genomicist at the National Human Genome Research Institute in Bethesda, Maryland. He and his colleagues compared DNA from 13 species, including the dog, cow, chicken, and puffer fish. In addition to genes, regulatory regions match, they report in the 14 August issue of Nature. And those regions, too, can confound gene counts.

One program, GENSCAN, stands out among the others as having set the standard for the field. When Chris Burge, now at the Massachusetts Institute of Technology in Cambridge, began writing the program in 1996, many of his colleagues were advocating a comparative approach. They picked out genes in a newly sequenced genome by matching its DNA against known genes in existing databases. But Burge disagreed. “We had human sequence, but there was really nothing to compare it to,” he recalls. No other vertebrate genomes were very far along, and the matches in sequenced genomes of the fruit fly, nematode, and microbes were fairly limited.

Instead, Burge took a lesson from David Haussler, a computer scientist at the University of California, Santa Cruz. Three years earlier, Haussler had realized that the gene-prediction problem was similar to the challenge faced by linguists who were trying to pick out patterns of syntax, grammar, and other features of languages. He and others suggested that their colleagues borrow a statistical tool from linguistics called a Hidden Markov Model. It calls for the program to make predictions based on a set of benchmarks it acquires from existing information.

View this table:

“There are a whole bunch of patterns and rules that distinguish parts of genes,” Burge points out. For example, all —or at least almost all—genes begin and end with a particular sequence. The ends of exons also have a characteristic sequence that tells enzymes to slice out the intron that follows. Burge “taught” the model these rules by having it analyze the sequences of several hundred genes with known intron and exon positions. The patterns it learned became the grist of a Hidden Markov Model for predicting whether a stretch of DNA includes a gene.

The approach proved a great success. Today, Hidden Markov Models are standard in most gene-prediction algorithms. As for GENSCAN itself, “it was significantly better than what was out there,” says Salzberg. Adds Brent, “It was just the best.”

But even the best have their flaws. GENSCAN's is overenthusiastic gene identification: It predicts 45,000 genes for the human genome, almost double the currently accepted total. Burge admits that GENSCAN has this problem but thinks that too many genes are better than too few; one can always eliminate the false positives.

GENSCAN will probably never predict the correct number of genes, Burge says. And much has changed in the genome world since the program was introduced. Taking account of new sequence data from humans and other species may be key to getting the gene tally just right. “If I were working on gene finding today, then the comparative approach would be the way to go,” he says.

Several programs, such as the official GeneSweep counter, Ensembl/Genewise, pick out genes based on their resemblance to what's already known. But they are much more sophisticated than earlier comparative efforts. Genewise, developed by Birney and his colleagues, works backward from known proteins. Proteins come in families whose members' amino acid sequences, and consequently DNA sequences, look more or less alike. Taking advantage of these family resemblances, the computer program compares new protein sequences derived from genes to previously discovered proteins from the same or different organisms.

Matching entire genomes rather than comparing short stretches of sequence is becoming more feasible and fruitful as more genomes are sequenced. Together, these comparative approaches are the most promising route right now, Affymetrix's Kulp says. And several programmers are melding multiple gene-prediction strategies.

Despite these advances, “we have probably reached a plateau,” Guigó complains. Few of the next-generation programs come up with similar gene totals. And dark matter still looms as a big unknown, one that no current program can touch. To decipher it, what's needed is knowledge about why genes have the characteristics they do and what dark matter genes might look like. In short, Guigó says, “to get better we will need to better understand the biology.” Nobody is betting on when that will happen.