# News this Week

Science  24 Apr 2009:
Vol. 324, Issue 5926, pp. 446
1. Biomedical Policy

# Draft Stem Cell Guidelines Please Many, Disappoint Some

1. Constance Holden,
2. Jocelyn Kaiser

They are not perfect, but they're a big improvement over what scientists have been living with since 2001. That's how most feel about the draft guidelines on human stem cell research released last week by the U.S. National Institutes of Health (NIH).

The proposed rules, to be finalized this summer, will expand the number of human embryonic stem (hES) cell lines available to researchers by eliminating the cutoff date for cell lines that qualify for federal funding. Acting NIH Director Raynard Kington predicted that “in a matter of months, we are likely to increase greatly the number of human embryonic stem cell lines eligible for federal funding.” He said that NIH estimates up to 700 lines exist based on literature reviews and registries.

But some restrictions remain. The cell lines must be derived from surplus embryos donated by couples receiving fertility treatment. Stem cells lines derived through research cloning, or somatic cell nuclear transfer (SCNT), which many researchers are eager to try, will not be eligible.

Some researchers are also concerned about just how many of the existing hES cell lines will be eligible, given NIH's newly detailed requirements on informed consent. But most share the sentiments of stem cell researcher Sean Morrison of the University of Michigan Medical School in Ann Arbor, who says the proposed policy is “a huge advance” over the Bush policy and “a reasonable compromise based on where the science stands.”

On 9 August 2001, then-President George W. Bush decreed that federal funding be restricted to research on stem cell lines derived before that date. To the delight of many scientists, President Barack Obama's 9 March executive order lifted those restrictions. But the order did not spell out the source of the embryos, leaving that to NIH to decide (Science, 20 March, p. 1552). Now researchers have the answer: The stem cells must be “derived from embryos created by in vitro fertilization (IVF) for reproductive purposes and were no longer needed for that purpose,” NIH's draft states. The use of cells from “other sources, including [SCNT], parthenogenesis, and/or IVF embryos created for research purposes, is not allowed.” It also notes that Congress has barred NIH from funding the derivation of stem cells from human embryos under the so-called Dickey-Wicker amendment.

Kington explained to reporters last week that in drafting its guidelines, NIH heeded both legal restrictions and professional guidelines, including a 2005 report on stem cell research from the National Academies. “There's strong, broad support” for allowing research on surplus embryos from fertility clinics, he said, as shown by legislation that passed Congress twice (and was twice vetoed by Bush). “There's not similar broad support for using the other sources.” Moreover, Kington pointed out, cell lines created solely for research either by IVF or SCNT don't yet exist as far as NIH is aware. NIH will review its policy as the science advances, he added. Funding continues to be allowed for research with induced pluripotent stem (iPS) cells—cultivated from adult cells—which many think will offer the same promise as cells from SCNT.

Many researchers would like to be able to work on hES cell lines derived in other ways—in particular, through SCNT. That technique might be used, for example, to make an embryo by inserting DNA from a skin cell of a patient with a disease into an enucleated egg. Researchers could then derive stem cells for studying that disease in test tube experiments.

Stanford University School of Medicine stem cell researcher Irving Weissman says the proposed ban on SCNT goes against the policy implied by Obama's earlier comments. “The NIH has not served its president well,” Weissman said in a statement. He said there is no prohibition on SCNT in guidelines established by the International Society for Stem Cell Research (ISSCR) or by the National Academies.

It's unclear how many more lines will qualify under the new rules. A limiting factor could be the informed consent procedures. NIH hewed closely to those recommended by the National Academies and ISSCR. But Harvard University stem cell researcher George Daley points out that some of the 21 lines eligible for federal funding under the Bush policy would not qualify under the “exhaustive informed consent language” that was “not widely practiced before 2006.” Therefore, there is some talk both at NIH and in Congress about the potential need to “grandfather” in some Bush-approved lines currently in use. (Grant applications already submitted to NIH will also need to pass muster.)

The draft policy includes a few other restrictions. NIH will not fund work that involves the possible introduction of pluripotent human cells (either iPS cells or ES cells) into the germ lines of any animals, a restriction recommended by the academies' report. Parthenotes, which are short-lived embryos created from an unfertilized egg, are also forbidden, as they qualify as human embryos under the Dickey-Wicker amendment.

The public has until 30 days after the policy is published this week in the Federal Register to submit comments, and NIH must issue final rules by 7 July. Some lawmakers have already weighed in. Representatives Michael Castle (R—DE) and Diana DeGette (D—CO), co-authors of the legislation that Bush vetoed, say they favor “more expansive guidelines”—a hint that an effort may be afoot to overthrow Dickey-Wicker. But Tom Harkin (D—IA), a Senate sponsor of the vetoed legislation, said he is “very pleased” with the NIH draft.

2. Archaeology

# Did Humans Learn From Hobbits?

1. Elizabeth Culotta

Thousands of small, sharp-edged flakes of volcanic tuff and chert have been unearthed from the cave of the “hobbit,” the roughly 1-meter-tall ancient human found on the island of Flores in Indonesia. The stone tools have puzzled researchers almost as much as the bones of the diminutive hobbits themselves. How could a hominid with a brain the size of a large pear craft tools? Now a detailed new analysis sheds light on the hobbit's technological capabilities and raises a new mystery: Why did the modern humans who arrived later on Flores make tools the same way hobbits did?

Archaeologist Mark Moore of the University of New England in Armidale, Australia, and his colleagues have been studying tools recovered from Liang Bua Cave on Flores since 2002. The most complete hobbit skeleton dates to about 17,000 years ago, but excavators claim other bones and tools date as far back as 95,000 years ago. All were excavated from below a layer of volcanic tuff dated to about 12,000 years ago. Above that layer, at 11,000 years and younger, researchers have found Holocene burials of Homo sapiens—modern humans—along with more tools.

Moore's team set out to analyze how the tools throughout the cave were made. The team reconstructed the sequence of blows struck thousands of years ago by analyzing the shapes of flakes, the position of scars left when flakes were struck off a stone, and other details. In a paper now in press at the Journal of Human Evolution and in a talk this week at a hobbit-themed meeting,* they report their analysis of 11,667 tools from five layers in the cave.

Moore found that hominids knapped stone in the same simple way throughout the 95,000 years represented in the cave layers they studied. Hominids first struck flakes off cores outside the cave. Then they brought large flakes into the cave to make smaller flakes, striking them directly with a hammer-stone or using an anvil. They used a handful of simple knapping techniques in repeated patterns that changed little over time. For example, tools in both the upper and the lower layers of the cave averaged about nine blows per tool.

Tools in the younger, H. sapiens sediments tended to be made of higher quality raw materials; 60% in that layer were chert, compared with only 17% in the lower layers. And more of the Holocene tools were burned, apparently from being left near a fire. But they were constructed in the same way as those from lower layers.

To Moore, the “most parsimonious” explanation is that the hobbit, H. floresiensis, made the older tools, some found in layers with their bones. Later, H. sapiens arrived on the scene and made similar tools. Moore even suggests that there may have been contact between the species, with modern humans copying H. floresiensis toolmakers before they went extinct. “The striking thing to me is the degree of similarity in the various permutations [combinations of techniques]. I can see how different hominins might converge on the techniques themselves, but I find it more difficult to understand how those permutations could be so similar without more direct observation or interaction.”

The analysis gets very high marks from other archaeologists. “They took a very close look at these materials, and I'd trust it,” says Nicholas Toth of Indiana University, Bloomington. “A model study,” pronounces archaeologist John Shea of Stony Brook University in New York state. Most important, the paper also “corrects an earlier impression that these are really sophisticated tools,” adds Alison Brooks of George Washington University in Washington, D.C. Some early reports had implied that the Liang Bua tools were sophisticated like those of modern humans—a finding at odds with the hobbit's tiny brain. But many archaeologists think even small-brained hominids could make this kind of simple tool, which resembles much more ancient tools from Olduvai Gorge. In fact, Toth and Indiana colleague Kathy Schick have trained bonobos to make similar cutting-edge tools.

However, when it comes to humans imitating hobbits, many archaeologists aren't yet buying. Oldowan-type technology is so simple that modern humans could have converged on the toolmaking pattern because of environmental needs, Brooks and others said. More fundamentally, with a suite of similar tools throughout the cave layers, “Homo sapiens might have made the lot of them,” says Shea. He points out that modern humans were moving through the Indonesian archipelago starting about 45,000 years ago and perhaps earlier. Toth agrees that when it comes to the identity of the tool-makers, “the jury is still out.” Moore counters that some Liang Bua tools are about 100,000 years old—meaning they were made long before there's any sign of our species in the region. Having H. sapiens craft the older tools might be as surprising as humans copying hobbits.

• * “Hobbits in the Haystack: Homo floresiensis and Human Evolution,” Stony Brook University, New York, 21 April 2009.

3. Biomedical Research

# Genome Scans: Impatient for the Payoff

1. Robert Koenig

A simmering debate in the genomics community about research strategy went public last week in commentaries published online by the New England Journal of Medicine (NEJM). A central disagreement concerns the likely value of chip-based genome scans to gauge inherited risks of developing common diseases. The discussion comes at a time when the U.S. National Human Genome Research Institute (NHGRI) is asking scientists to help set its course in a new 5-year road map.

In the most provocative of four NEJM articles, David B. Goldstein, director of the Center for Human Genome Variation at Duke University in Durham, North Carolina, says the first 100 or so genome-wide association studies (GWAS), which use gene chips to find associations between common gene variants and diseases, have identified important variants that appear to influence disease risk, but the impact of most of those variants is relatively low. He told Science that, once such association studies “have been run on the first few thousand patients for a given disease, there is only marginal return in pushing the sample sizes up further.” Instead, he wants to shift more research to full sequencing of patients' genomes to find “rarer variants of larger effect” linked to disease.

Other scientists, while agreeing with parts of Goldstein's commentary, argue that GWAS will be highly useful; they have already found more than 200 loci associated with certain diseases, even though most raise disease risks by relatively small amounts. They argue that improved chips may yet reveal a new cache of less common genetic variants that have a large effect on disease. Many also favor sequencing complete patient genomes, but only after the cost comes down.

The divergent perspectives about GWAS present “a classic example of the glass being half-full or half-empty,” says Francis Collins, NHGRI's former director, who is writing a book on personalized medicine. “On the half-full side, researchers have now discovered a stunning number of genetic risk factors for common disease, after many years of frustration. While for the most part, those risk factors have only modest odds ratios, they are pointing us toward biological pathways that we desperately needed to understand.” From the “half-empty” perspective, Collins says, “it is clear that most of the heritability for common disease has not yet been discovered—this is now generally referred to as ‘the dark matter of the genome.’”

That dark matter is at the center of the current genomics debate. As part of a long-range planning effort, NHGRI is soliciting suggestions from scientists about the best ways to apply genomics to clinical problems and set the most productive course for genome sequencing. Teri Manolio, who directs NHGRI's Office of Population Genomics, says, “We plan to move in the direction that the science takes us,” but strongly defends the results from the GWAS approach because it is identifying critical biological pathways involved in disease.

David Altshuler, who co-chairs the international consortium for the 1000 Genomes Project, which aims to sequence at low coverage the genomes of 1000 individuals, says that the massive new genome catalog he's helping to create will extend the reach of GWAS. The project will capture single-nucleotide polymorphisms (SNPs) that are rarer than those now available—including those present in 1% or more of the population—making it possible to probe genomes more completely (Science, 25 January 2008, p. 395). The consortium includes the Wellcome Trust Sanger Institute in the United Kingdom and the Beijing Genomics Institute in Shenzhen, China. Altshuler, with appointments at the Broad Institute in Cambridge, Massachusetts, and Massachusetts General Hospital in Boston, says the 1000 Genomes group aims to produce its first paper at the end of this year and a final publication by the end of 2010. While Collins hopes that the 1000 Genomes Project will help “uncover a whole new potential cache of less common variants of large effect,” he cautions that those variants “will have to be tested in appropriate case-control studies, since there is no phenotypic information available on the 1000 Genomes DNA samples.”

Judy H. Cho, genetics director of Yale University's Inflammatory Bowel Disease Center, who has helped identify some of the 35 or so genes now associated with Crohn's disease, says the GWAS approach “has been a victim of its own success. Four years ago, the concern was: Will we find anything? But now critics ask: Why didn't you find everything?”

Researchers contacted by Science agreed that far more work needs to be done before physicians can make good use of personal genomic assessments. Peter Kraft, a Harvard School of Public Health epidemiologist who co-authored with Harvard colleague David J. Hunter an NEJM commentary on the current limitations of genetic risk prediction, says, “If we go forward doing GWAS studies as we are now, testing for inherited susceptibility on the basis of common risk alleles may be more effective in 2 or 3 years.”

Although it has become the focus of much public attention, predicting personal risk is not the point of GWAS, argues Joel N. Hirschhorn, a genomics researcher at the Broad Institute and Harvard University Medical School who wrote an NEJM commentary that offered a positive assessment of GWAS results. “The goal is not individual risk analysis but rather discovering the biological pathways underlying diseases,” he says.

While he takes issue with Goldstein's overall assessment of GWAS, Collins agrees that “less common variants of larger effect … are going to account for a lot of the missing heritability, though copy-number variants will also likely play a role. To get those answers, we need to shift attention from SNP chips to actual sequencing.” But despite recent advances in technologies, several scientists say, the full sequence strategy still looks expensive for now.

4. Newsmaker Interview

# Brilliant Moves to Tackle Global Threats

1. Greg Miller

Tackling big problems has been a lifetime pursuit for Larry Brilliant. A physician and epidemiologist, Brilliant played a key role in the World Health Organization's smallpox eradication program in the 1970s. More recently, he directed Google.org, the Internet company's philanthropic arm. In February, Brilliant stepped down from that role—amicably, he says—as Google narrowed its philanthropic focus to emphasize projects that tap into its strength in technology, such as Google Flu Trends, which maps flu activity based on online searches for flu-related terms. Brilliant stayed on as Google's Chief Philanthropy Evangelist but last week announced that he's leaving for a brand-new venture, a nonprofit foundation started by Jeffrey Skoll, the billionaire founding president of eBay.

Skoll has kicked in $100 million in seed money with a promise of more for the Skoll Urgent Threats Fund, which will tackle what he sees as five of the most urgent threats facing the planet: climate change, water scarcity, pandemics, nuclear proliferation, and conflict in the Middle East. As the organization's president, Brilliant's job is to make sure the money is well spent. Brilliant spoke with Science on 16 April. His comments have been edited for brevity. Q:What's the idea here? L.B.:The Earth has a number of urgent threats that are interwoven. If there were a single nuclear explosion in anger, all the work that we're doing on climate change, on trying to bridge the gap between rich and poor … would be set back generations. If any one of these urgent threats were to occur [or escalate], the work that we do would be so much more difficult. They have to be prevented or mitigated. Q:How do you hope to accomplish that? L.B.:We'll be trying to create the public will and confidence to tackle these seemingly impossible challenges. We'll be doing as much advocacy as grantmaking. Having the media savvy of Participant [Media, Skoll's production company, which has bankrolled socially conscious films such as Fast Food Nation and An Inconvenient Truth] was a huge draw for me. Jeff has learned how to stiffen the will of a community to fight the good fight. Q:Will you be funding any research? L.B.:Absolutely. I'm looking for research that changes people's minds. It's research that lets people know what you can do. In smallpox, when Bill Foege wrote his paper on selective epidemic control, we knew that by vaccinating 100 people—the right 100 instead of the wrong 100 million—you could extinguish the epidemic. I'm very interested in digital surveillance technologies. In the area of drought, I'm very interested in ways to move water around. We've got plenty of water on the planet; we just don't have it where people live. Each of these different issues requires creative thinking and research to support breakthrough ideas that help change people's pessimism. Q:What else? L.B.:We need scientists for the next millennium. We don't have them. We don't have the half—genomic molecular biologist/half-epidemiologist and other dual-trained, cross-trained individuals or departments that are going to help us do the genomic epidemiology that will stop the next virus from jumping from an animal to a human and creating a pandemic. We are lacking scientific workforce for the 21st century because the problems of the 21st century have grown faster than the ability of training programs to keep up. 5. ScienceNOW.org # From Science's Online Daily News Site Closer Look at Einstein's Brain. When a rare genius like Albert Einstein comes along, scientists naturally wonder if he had something special between his ears. The latest study of Einstein's brain, published in Frontiers in Evolutionary Neuroscience, concludes that certain parts of it were indeed very unusual and might explain how he was able to go where no physicist had gone before when he devised the theory of relativity and other groundbreaking insights. http://tiny.cc/F1MtB Neandertal Babies Didn't Do the Twist. Giving birth is more difficult—and dangerous—for modern humans than for any other primate. Not only do human mothers have to push out babies with unusually big heads, but infants also have to rotate to fit their heads through the narrow birth canal. Now, a new virtual reconstruction of the pelvis of a Neandertal woman published in the Proceedings of the National Academy of Sciences suggests that Neandertal mothers also had a tough time giving birth to their big-headed infants—but the babies, at least, didn't have to rotate to get out. http://tiny.cc/Z5yEb Red Leaves Say, “Bug Off!” Come autumn, leaves exchange their lush greens for deep reds, but why? Scientists have puzzled over this transformation for more than a century. Now a new study of aphids and apple trees in Proceedings of the Royal Society B reinforces the theory that the red hue wards off insects looking for a leafy snack or a place to nest. http://tiny.cc/dPeBA The Mother of All Pileups. In a faraway corner of the universe, a crash of cosmic proportions is under way, cramming more than 1000 galaxies into a space normally reserved for a handful. Astronomers studying the phenomenon report in The Astrophysical Journal that what they learn should improve their understanding about how the largest structures in the cosmos have evolved. http://tiny.cc/WqT5g Read the full postings, comments, and more at sciencenow.sciencemag.org. 6. Scientific Misconduct # Science Retracts Discredited Paper; Bitter Patent Dispute Continues 1. Dennis Normile In an unusual decision, Science this week retracted a 2005 report without the agreement of all the authors (see p. 463). The report, authored by a group at the Korea Advanced Institute of Science and Technology (KAIST) and at CGK Co., both in Daejeon, South Korea, describes a method, dubbed MAGIC, to identify drug targets by tracking protein movements in live cells (Science, 1 July 2005, p. 121). The retraction is based on an investigation by KAIST that concluded that although the technique might be valid, data in the paper were fabricated and “the extent of the fabrication is serious enough to damage the authenticity of the entire paper.” The same group claimed in the journal Nature Chemical Biology in July 2006 that it had used MAGIC to identify an antiaging molecule; this report was retracted last July. But in a twist, a legal battle is raging over the intellectual property at the heart of the discredited papers. “We strongly believe that there is value in the patent rights to the MAGIC technology,” KAIST's Research Integrity Committee wrote in an e-mail responding to questions from a Science reporter. “Even though the data and results in the paper were fabricated, the idea of the methodology is original.” And the senior author of both retracted papers staunchly defends the findings. “A number of strong evidences will be provided to prove that MAGIC technology and anti-aging compounds in two papers are real,” chemical geneticist Tae Kook Kim, formerly at KAIST, wrote in an e-mail to a Science reporter. He added that he intends to take legal action “against several parties for my defamation and libel.” The convoluted saga began in July 2004 when Kim and several partners established CGK Co. to commercialize a technique for identifying drug targets. The method, MAGnetism-based Interaction Capture, works by coating a magnetized nanoparticle with a molecule of interest. The coated nanoparticle is then introduced into a cell in which a target protein has been tagged with a fluorescent label. Applying a magnetic field forces the nanoparticle to move. If the fluorescence moves in concert, that indicates that the molecule of interest has bound to the target protein. Kim and two members of his lab, Jaejoon Won and Yong-Weon Yi, are listed as inventors in a patent application, according to a translation of a Korean Intellectual Property Office document that CGK provided to a Science reporter. After describing MAGIC in the Science paper, the group reported in Nature Chemical Biology how the technique had been used to identify a molecule, CGK733, that resets a cell's intrinsic aging clock by inhibiting a certain protein. Won was first author and Kim was corresponding author on both papers. In July 2006, the same month the report appeared in Nature Chemical Biology, CGK completed raising more than$2.5 million in venture capital, according to a statement CGK provided to a Science reporter on 10 March 2008.

But that statement also notes that CGK had trouble getting MAGIC to work. This led Yi—an author on both papers who by then had moved from KAIST to CGK to become the company's chief technology officer—to ask Science and Nature Chemical Biology to remove his name from the papers in December 2007, according to the CGK statement. On 11 February 2008, CGK informed KAIST of concerns about MAGIC. KAIST launched its own investigation on 12 February 2008.

On 28 February, the university informed both journals that although its investigation was continuing, the panel had come to a preliminary conclusion that “the two papers do not contain any scientific truth,” according to an “editorial expression of concern” Science posted on its Web site on 3 March 2008 (Science, 14 March 2008, p. 1468). On 13 March, KAIST issued a press release in English summarizing the results of the preliminary investigation and alleging that Won, Yi, and Kim were involved in or aware of the misconduct. Unlike the press release, the final report later shared with Science does not assign responsibility for alleged misconduct.

In the e-mail to the Science reporter, Yi says he played a minor role in the project and even questioned Kim about his inclusion as an author “because I thought I hadn't contributed enough.” Before the investigation, he claims, he hadn't suspected superiors would commit misconduct: “It was almost impossible for me to find out the truth” from them, he wrote.

In its July 2008 issue, Nature Chemical Biology published a letter signed by eight of the nine authors retracting the Won et al. paper. The letter states that the preliminary KAIST investigation “revealed several irregularities.” One was that the “application of MAGIC technology for identifying ATM (a key protein) as the target of CGK733 was fabricated.” It also says “our original notebooks and data are not available to substantiate the scientific claims of the paper.” Kim did not sign the letter. An editor's note reads: “T.K. Kim supports the retraction of the paper but maintains that the irregularities are confined to Figure 2 of the paper—specifically, that the MAGIC screening was improperly performed and the chemical structure of CGK733 was misrepresented.”

Retraction of the Science paper did not go as smoothly. Science requires that all authors agree to a retraction, says Monica Bradford, Science's executive editor. But the journal could not reach one co-author, Neoncheol Jung, whose affiliation was listed as CGK. The whereabouts of Jung, who was CGK's founding CEO, remain unknown. Yeon-Soo Seo, a KAIST biochemist who served on the investigating committee, says his panel received one e-mail from Jung, who wrote that he was not involved in the case, he had nothing to do with the fabrication, and his name was included in the paper against his will. Lacking Jung's agreement and without full consent of authors and their institutions, Bradford says Science waited for KAIST to provide the investigating committee's final report, which had been completed in June 2008. But the report's release, she says, “got tied up with a legal dispute.”

The legal dispute is murky as well. CGK claims it reached an agreement in August 2004 with KAIST assigning intellectual property and commercial rights to the MAGIC technology to the company. But in March 2007, KAIST initiated legal action to reclaim those rights, according to statements from both CGK and KAIST. “KAIST hoped that through those legal proceedings they would have a chance to learn more from CGK,” Bradford says she was told by a KAIST official. CGK, however, “was pushing for retraction,” she says. CGK sent a copy of what it said is the KAIST investigating committee's final report, dated 22 May 2008, to Science editors, who had it translated. Science delayed taking action until it received an official version from KAIST, Bradford says.

Meantime, CGK devised an alternative to MAGIC, described by its researchers in the 10 December 2008 issue of the Journal of the American Chemical Society. CGK has applied for two patents, one of which has been granted, related to the technology, which, like MAGIC, relies on coated magnetic nanoparticles and fluorescent-tagged proteins. In an e-mail to a Science reporter, CGK charged that KAIST had delayed the retraction process to bolster its legal position; KAIST strongly denied this in an e-mail.

Although a Korean court on 9 December returned MAGIC patent rights to KAIST, the Korean Intellectual Property Office recommended on 19 January that the patent be rejected. The final decision is expected this summer; if it goes against KAIST, KAIST officials have said the company will appeal.

On 28 February 2009, KAIST forwarded to Science parts of the final report related to the Science paper. The authors could not provide notebooks or original data for any of the experiments described in the paper, according to a translation of the report prepared for Science. The KAIST report notes that Won and Tae Kook Kim admitted that experiments supporting Figure 2 in the Science paper, which purports to show magnetic manipulation of proteins tagged with fluorescent markers, were not carried out as reported. Based on the KAIST report, “The data, results, and conclusions in the Won et al. report are clearly not reliable,” writes Science Editor-in-Chief Bruce Alberts.

Bradford says that when she informed senior author Tae Kook Kim that the paper would be retracted, he said he would agree if the wording was changed to indicate that only parts of the paper were not reliable. Science declined. “Our paper, as it stands, cannot be substantiated; it's got to go,” Bradford says.

Repercussions for the Korean scientists have been severe. Last November, KAIST dismissed Tae Kook Kim; CGK is suing him for criminal fraud, says CGK's Dae-Joong Kim. Sometime after the papers were published but before problems arose, Won took an “associate specialist” position at the University of California, Los Angeles; he left that post in July 2008, according to a UCLA spokesperson who declined to describe the circumstances, citing privacy concerns. Won did not respond to an e-mail or to a voice mail message seeking comment.

Late last year, Yi sought to have criminal defamation charges brought against five members of KAIST's investigating committee, according to Yi's e-mail. Prosecutors declined to pursue criminal charges on 29 December, but Yi has appealed to a court.

The KAIST Research Integrity Committee says it “learned a lot from this incident” and is planning educational programs “to prevent research misconduct and to promote research ethics and integrity.”

7. ScienceInsider

# From the Science Policy Blog

An innovative approach to provide the best malaria drugs to the world's poor officially got under way last week in Oslo. Instead of providing money to governments to buy the medicines for their public health systems, the Affordable Medicines Facility for malaria (AMFm) (Science, 21 November 2008, p. 1174) will subsidize companies to sell the drugs on the private market at bargain prices. Most of the world's poorest rely on small pharmacies for their drugs, but to save money, they often pick cheap, ineffective drugs or counterfeits instead of state-of-the-art combination therapies. So far, AMFm has received $225 million to tackle the problem. The U.S. Environmental Protection Agency has asked scientists how to revise the Clean Water Act to protect seas against ocean acidification from atmospheric carbon dioxide. Under the current rules, waters are designated as impaired if their pH deviates from naturally occurring levels by 0.2 units. But biologists say that some organisms are affected by smaller changes. A more complex approach would also take into account how organisms or ecosystems are affected differently by changing pH levels. Last week's announcement that William Brinkman, former head of research at Bell Labs, would be nominated to run the$4.8 billion Office of Science at the Department of Energy suggests that Energy Secretary Steven Chu, another alum, hopes to tap the fabled lab's expertise at marrying basic and applied research.

Elsewhere … Bush science adviser John Marburger returned to Washington, D.C., to make science policy more scientific. The Union of Concerned Scientists released a report suggesting that genetically engineered crops have not outperformed more traditional varieties. British spy service agency MI5 is looking for a scientific adviser—the equivalent of James Bond's Q. Time to spell check your résumé.

For the full postings and more, go to blogs.sciencemag.org/scienceinsider.

8. German Physical Society

# A New Take on Doping in Iron-Based Superconductors

A year ago, physicists discovered a new family of high-temperature superconductors: materials that carry electricity without resistance at temperatures as yet inexplicably far above absolute zero. Immediately, researchers wondered whether the compounds, which contain planes of iron and arsenic atoms, work in the same way as the only other high-temperature superconductors, the copper-and-oxygen, or “cuprate,” compounds that have defied explanation since their discovery in 1986 (Science, 16 May 2008, p. 870). New data suggest that a key step in making all the superconductors—“doping” them with impurities—plays a very different role in each of the two families.

In the cuprates, doping produces superconductivity by adjusting the number of electrons in the current-carrying copper-and-oxygen planes. But in the iron-and-arsenic compounds, doping appears to induce superconductivity by altering the material's crystal structure, report Helge Rosner and colleagues at the Max Planck Institute for Chemical Physics of Solids in Dresden. The result jibes with other evidence, says Dirk Johrendt, a chemist at Ludwig Maximilians University in Munich, who was not involved in the study. “One has to look to see if some sort of more subtle structural change is involved,” he says.

In a cuprate such as strontium-doped lanthanum copper oxide, the lanthanum atoms lie between the copper-and-oxygen planes, in which the atoms form a pattern of squares. Without the strontium, the mobile electrons sit one to each copper atom and push against one another so hard that they get stuck in a giant traffic jam. The strontium atoms, which replace some of the lanthanum atoms, soak up enough electrons to break the impasse. The electrons then pair—in a way not yet fully understood—to flow without resistance.

Things seem to work similarly in an iron-based superconductor. For example, strontium iron arsenide is an ordinary metal. Replace enough of the strontium with potassium, however, and superconductivity appears at temperatures below 38 kelvin. But that's not necessarily because the number of electrons changes.

To make that case, Rosner and colleagues left the strontium alone and instead replaced some of the iron with cobalt, an atom of almost the same size but with one more loosely bound electron. Such doping increased the density of mobile electrons but still induced superconductivity below 20 kelvin. The researchers also replaced some of the iron atoms with ruthenium atoms, which have as many loosely bound electrons but are bigger and put a larger strain on the crystal structure. Again, superconductivity emerged.

Those results suggest that changes in the material's crystal structure are the key. Physicists know that, at a temperature above the superconducting transition, an iron-and-arsenic superconductor undergoes a structural change in which the squares in the iron-and-arsenic planes squash slightly into a diamond shape. Neighboring rows of iron atoms also become magnetized in opposite directions in a pattern called “antiferromagnetism.” The usual doping suppresses this transition and produces superconductivity. But so does the doping that leaves the electron density unchanged, suggesting that “you can get the same effect by making the structural change only,” Rosner says.

The structural changes may be all that's relevant in the iron-and-arsenic compounds, Johrendt says. Rosner hesitates to go that far. In any case, such experiments underscore a difference between the iron-and-arsenic superconductors and the cuprates: Replace copper atoms in a cuprate's planes with anything else, and the material's superconductivity vanishes.

9. German Physical Society

# Leverage: The Root of All Financial Turmoil

During the 1990s, physicists flocked to Wall Street and other financial hubs, eager to turn their analytical skills and phenomenological mindset to the problem of making a killing. Now that the world's stock markets are in retreat, they've turned to explaining why markets crash. According to one new analysis, leverage—the practice by hedge funds and other investors of borrowing money to buy investments—is the root of many nettlesome properties of financial markets that classical economics cannot explain, including a propensity to crash.

Given that in the buildup to the recent global economic meltdown hedge funds had been leveraging their deals by ratios of 30-to-1 (that is, borrowing $30 for every$1 of their own that they put in), it may seem obvious that massive leverage leads to trouble. But Stefan Thurner, an econophysicist and director of the complex systems research group at the Medical University of Vienna, Austria, and colleagues say their model shows that many of the distinctive statistical properties of financial markets emerge together as rates of leverage climb. “Leverage is the driver,” Thurner says. “That wasn't obvious.”

Financial markets behave in ways that, econophysicists say, classical economic theory cannot explain. Classical economics assumes that the fluctuations in stock prices conform to a so-called Gaussian distribution—a bell curve that gives little probability to large swings. In reality, the distribution has “fat tails” that make big changes more likely, and the shapes of those tails conform to a mathematical formula known as a power law. Classical economics assumes that the fluctuations are uncorrelated from one moment to the next, whereas big swings in prices tend to come together in the so-called clustering of volatility.

To try to explain those characteristics, over the past 5 years Thurner and colleagues have developed an “agent-based model” of a market. In such a computer model, virtual agents of various types interact according to certain rules, like robots playing a game. The researchers included hedge funds that could borrow to make their investments; banks to loan the money; “noise investors” who, like day traders, simply react to the market and have no other insight into the value of assets; and general investors who played the role of, for example, state pension funds.

The model contains more than a dozen adjustable parameters. However, Thurner and colleagues found that the maximum level of leverage exerts a curious, unifying effect. If they forbade leverage, the market behaved largely as classical economics would predict. But as they increased the maximum leverage, the characteristics of real markets emerged together. “We can explain the fat tails, the right [power law], the clustering of volatility, all this,” Thurner says. And when the leverage limit climbed to levels of 5-to-1 and beyond, the market became unstable and hedge funds went bust much more often.

“I thought it was rather brilliant,” says Christoph Jan Hamer, an econophysicist with Solvency Fabrik in Köln, Germany. Hamer says he was impressed with a detail of the model: If leverage is high, then a tiny fluctuation created by the noise traders can trigger a much bigger swing. But Christian Hirtreiter of the University of Regensburg says, “I would think that leverage itself is not the problem. I would think it is a symptom of the problem.”

Thurner, who managed a hedge fund that tanked, says that limiting leverage should help prevent crashes. He admits, however, that he would not have embraced that idea when the market was still going strong.

10. German Physical Society

# Snapshots from the Meeting

Piping in nanotubes. Carbon nanotubes could be made into minuscule electrical and mechanical devices on microchips—if researchers could put the things where they want them. Pushing tubes around with a tiny fingerlike probe is slow; depositing them in a solution is relatively imprecise. To do better, Sergei Loschek of the Chemnitz University of Technology in Germany and colleagues carve in a layer of polymer microfluidic channels to pipe the solution where they want it and rely on electric fields to pull the tubes into place. The polymer can be laid down and removed in minutes, Loschek says, making the method fast and precise.

A surprisingly squishy virus. Many viruses possess a hard shell of pure protein, but the influenza virus packs itself into a container with an inner layer of protein and an outer layer of fatty molecules called lipids. That shell is no harder than the lipid layer alone, as Iwan Schaap and colleagues at the British National Institute for Medical Research in London have shown by squeezing the viruses with a microscopic probe. The proteins in influenza's shell “are definitely not there to reinforce the virus,” Schaap says. The question is, What purpose do they serve?

Back to school on university assignments. Four years ago, Germany scrapped its centralized system for assigning students to universities for competitive programs such as medicine. Now, students apply to individual universities. The best get accepted by all their choices whereas others get stuck on waiting lists—sometimes for half a year. But simulations by Christian Hirtreiter of the University of Regensburg and colleagues suggest that the old system placed as many students at their preferred schools. Moreover, they say, they can optimize the old system to cut by half the several percent of students whom it did not assign to one of their first-choice schools.

11. German Physical Society

# Water Droplets Grow Faster Than Expected

Breathe on a pane of glass and droplets of water will condense on it. Physicists have analyzed that bit of everyday physics and have broken it down into three stages. First, tiny individual droplets form, or “nucleate.” Then, the droplets grow. Finally, they bump into each other and merge. Assuming that moisture condenses out of the air at a constant rate, during the second phase, the volume of each individual droplet should increase linearly with time—that is, increasing by the same amount each second.

But is it so simple? Mordechai Sokuler of the Max Planck Institute for Polymer Research in Mainz, Germany, and colleagues have shown that, paradoxically, whereas droplets in a crowd do grow steadily with time, an isolated droplet grows faster as it gets bigger: Its volume increases in proportion to time raised to the 3/2 power.

To prove it, Sokuler and colleagues performed one of the simpler of the thousands of experiments described at the meeting, growing drops in isolation or in groups on a small finger of glass in a chamber in which they could control the temperature and humidity. They then filmed the droplets condensing or evaporating. “I'm really glad that I got to do this project because it's so simple,” Sokuler says. Nobody seems to have done the experiment before, he says.

At first blush, it seems contradictory that an isolated droplet would grow faster than one in a group, but the odd effect has an explanation, Sokuler reports. The growth rate depends on the amount of excess water vapor in the air. Above a single droplet, the vapor distribution conforms to the dome-shaped droplet, and the amount of vapor condensing into it at any moment increases with its radius. But when many droplets lie on a plane close together, each distorts the vapor distribution near its neighbors, effectively smoothing out the distribution across the plane. That makes all the droplets grow steadily in time, regardless of their size.

“His main argument was a change in geometry,” says Thomas Thurn-Albrecht of Martin Luther University of Halle-Wittenberg. “This seems convincing to me.” It's a simple effect, he says, “but sometimes to think of a simple experiment is an admirable thing.”

12. Archaeology

# A New Look at the Mayas' End

1. Heather Pringle*

Climate reaserchers have fingered drought in the collapse of the great Maya civilization, but many archaeologists say it doesn't fit their data. Ultralocal paleoclimate indicators may spark a resolution

In the early 8th century C.E., the Maya city-state of Tikal eclipsed all rivals, becoming the most populous polity in the Americas. As many as 62,000 Maya nobles, artisans, and others squeezed into Tikal's crowded residential districts in what is now Guatemala; another 30,000 people, according to one estimate, toiled in verdant cornfields nearby. Tikal's divine kings ruled in splendor. They gleamed with jade jewels, commanded an army of artists and scribes, and presided over the construction of monumental causeways, temples, and pyramids.

During the century that followed, however, Tikal fell on hard times. Its building boom collapsed, its artists ceased to carve hieroglyphic inscriptions and paint murals, and its kings vanished. After 830 C.E., Tikal's population plummeted to just 15% to 20% of its peak. And Tikal was not alone. Elsewhere in the Maya world—a 324,000-square-kilometer area spanning southeastern Mexico and upper Central America (see map, p. 455)—dozens of other city-states crumbled between 695 and 1050 C.E. The big question is why?

Archaeologists have proposed many theories, from peasant revolts to plagues and volcanic eruptions. But starting in the 1980s, many researchers focused on the role of climate. The idea gained traction in the 1990s, and today paleoclimatic indicators from lakes in the Yucatán Peninsula and the ocean floor off Venezuela suggest that a series of devastating droughts struck the Maya lands beginning in 760 C.E. The evidence for these climatic catastrophes “is overwhelming now,” says Richardson Gill, an independent archaeologist based in San Antonio, Texas. “How could 95% of the cities in the Maya lowland, which relied on surface water reservoirs that had to be replenished annually, survive a 9-year drought?”

Many other Maya archaeologists, however, say the megadrought theory just doesn't fit their findings. Some Maya centers fell before the proposed droughts began, whereas others flourished even during the parched times. Several cities in the humid lowlands failed before those in drier regions did. The wide variation in space and time of the collapse has left many climate scientists and archaeologists at loggerheads. “The idea that people were dying in the Maya plazas from thirst is all very overblown and over-sensationalized,” says archaeologist David Webster of Pennsylvania State University (PSU), University Park.

New studies, several of which are being presented at the Society for American Archaeology (SAA) meetings in Atlanta, Georgia, this week, suggest productive ways out of this impasse. Researchers are using novel paleoclimatic indicators tied closely to specific archaeological sites to see just what ancient Mayans experienced; others are modeling the climatic impact of deforestation. These efforts show a new willingness among archaeologists to work hand-in-glove with the paleoclimatologists, says archaeologist Brian Fagan, a professor emeritus at the University of California, Santa Barbara (UCSB). “This is a form of multidisciplinary research that we never ever envisioned before,” Fagan says. “It's very exciting.”

## The slow collapse

The Maya occupied one of the most environmentally varied territories in the world—a patchwork of coastal plains, scrub forest, tropical forest, and temperate highlands. Today, annual rainfall ranges from 500 millimeters in the north to 4000 mm in the central lowlands, but most rain falls from May to December. To obtain water the rest of the year, the Maya settled along rivers and lakes and built reservoirs and canals to manage rainwater. By the Late Classic period, from 600 to 900 C.E., the Maya had founded more than 100 urban centers.

The collapse of the city-states began about 695 C.E., in the wet Petexbatún region of southern Guatemala, and proceeded sporadically over the next 350 years. Given the protracted timetable and the clues unearthed at individual sites, many archaeologists concluded by the early 1990s that the Classic Maya civilization was felled by a toxic cocktail of social and environmental factors, including war, overpopulation, soil erosion, and restive populations tired of the demands of self-aggrandizing rulers. It's a view many archaeologists still hold. “When you have something as complex as Classic Maya society, it's going to take a complex string of events to bring that to an end,” says Andrew Scherer, an archaeologist at Baylor University in Waco, Texas.

But beginning in 1995, geologist David Hodell of the University of Cambridge in the United Kingdom and his team began pointing to the role of climate change. They analyzed gypsum—which accumulates as water evaporates in dry periods—in cores from a lake in northern Yucatán and showed that the period from 750 to 850 C.E. was the driest in a 7000-year-long period (Science, 18 May 2001, pp. 1293 and 1367). Other indicators supported this picture. A Swiss-led team studied titanium—which declines in river-borne sediments when water erosion decreases during parched periods—in a marine core from the Cariaco Basin off Venezuela. They found a dry period at 760 C.E. and then three severe multiyear droughts: 810, 860, and 910 C.E. (see graph, p. 456, and Science, 14 March 2003, p. 1731). Then in a 2007 paper in Palaeogeography, Palaeoclimatology, Palaeoecology, researchers analyzed a variety of data—oxygen isotopes for gauging rainfall, luminescence for studying the rate of water flow from a cave ceiling, and other indicators—from a stalagmite from the Macal Chasm cave in western Belize. They found four extreme dry periods centered at about 780, 910, 1074, and 1139 C.E.

Some archaeologists were convinced. But when others went looking for evidence of severe climate change at their sites, they could not find it. In the Petexbatún region, for example, a 7-year interdisciplinary project by Arthur Demarest of Vanderbilt University in Nashville, Tennessee, concluded that the inhabitants had a relatively stable diet of corn and meat throughout the collapse period. Skeletal remains displayed no increases in anemia or infectious disease, as might be expected in a starving populace. “If there's a drought, but it doesn't affect people's health, then what does it matter?” says Demarest. Instead, his team uncovered extensive evidence of a bitter, nearly century-long rivalry between the region's rulers over trade routes, which sparked intense warfare and the destruction of the site of Dos Pilas in 760 C.E.

One difficulty in correlating climatic and archaeological data at some sites is the relatively large error ranges of radiocarbon dates. Another may be regional variation: “Droughts can be quite localized,” notes Jason Yaeger, an archaeologist at the University of Wisconsin, Madison. “So I'm a little skeptical about taking the Venezuela record and applying it directly to the Maya lowlands without local proxy records.”

One way out is to devise paleoclimatic indicators that come directly from dated archaeological sites. That's just what zoo-archaeologists Kitty Emery and Erin Kennedy Thornton of the University of Florida, Gainesville, set out to do, by using the animal bones that litter Maya deposits to track the proportions of wetland fauna over time. The researchers “have hit on something really exciting here that no one else has done before for the Maya. And that is to use the habitat information of the animals hunted, exploited, and eaten by the Maya as a proxy for climate change,” says archaeologist Heather McKillop of Louisiana State University in Baton Rouge.

Emery and Thornton, who presented their unpublished study this week at SAA, compiled lists of animal bones from 22 dated Maya sites and selected 15,000 identified specimens from more than 65 species, including animals such as jaguars that sometimes hunt near water and storks and musk turtles that spend much of their lives in wetlands. Then the researchers gathered published habitat data on the proportion of time each animal spent in aquatic or wetland environments and verified this with ecologists. “We can say that a particular species will spend 20% its time in one habitat and 10% in another,” says Emery.

She averaged the values of all species to get pictures of available habitat at given points in time and looked for changes in the proportion of wetland animals from 1800 B.C.E. to 1821 C.E. (Examining shifts in the proportion of wetland animals, rather than absolute numbers, should correct for varying hunting intensity and for animal responses to habitat change, she said.)

Although she cautions that larger samples are needed, the results so far fit well with the regional paleoclimate data: Swamp-loving species at 22 sites increased from 1% to 7% of specimens from 600 to 800 C.E., a period other proxies pointed to as relatively moist. But during the Maya collapse from 800 to 1000 C.E., swampland animals declined in all five watersheds from 7% to 2%, strongly suggesting that wetlands had shrunk markedly. However, when viewed from this ultralocal perspective, the dry periods look serious but not catastrophic, says Emery. Not one wetland species in the study disappeared completely between 800 and 1000 C.E. “I don't think there was anything that killed off enough fauna or enough landscape that it would have caused a collapse of human populations,” she says.

## Appeasing the rain gods

In bark-paper books, Maya scribes painted pictures of a rain god they called Chac. Artworks show Chac residing in a cave and creating precipitation by pouring water from an overturned jar. Between 680 and 960 C.E., Maya in western Belize created a Chac-oriented drought cult, centered in caves, to appeal for rain, according to a study in the current issue of Latin American Antiquity by University of Arizona, Tucson, archaeologist Holley Moyes, Belize Institute of Archaeology president Jaime Awe, and their colleagues. At Chechem Ha Cave, the team mapped 300 meters of tunnels and excavated the central chamber. To illuminate the passages—which appear to have been used exclusively for ritual—the Maya carried wood torches that spewed charcoal, and Moyes's team plotted the specks on each excavated surface. The density pointed to the intensity of ritual use and to where ceremonies were performed. The charcoal provided 48 radiocarbon dates, showing that the Maya had visited the cave repeatedly from as early as 1300 B.C.E. to about 950 C.E. Over time, the Maya left behind 1901 ceramic shards or complete pots.

The team discerned a major shift in ritual practices. Before 680 C.E., the Maya journeyed to the cave's central chamber to perform rites around a giant stalactite and pool, sometimes leaving broken pottery and twice leaving an intact jar and whole dish. After 680 C.E., however, ritualists deposited 51 whole jars and numerous other partially intact vessels, sometimes inverted, throughout the cave, including areas previously unvisited. “They left more ceramics between 680 and 960 C.E. than all the other periods put together,” says Moyes. Many of the vessels were large, wide-mouthed jars likely used for water collection. “These jars are what rain deities use,” says Moyes. “So this is the perfect gift for a rain deity.”

The changing rituals occurred during a prolonged dry period in the region, from 700 to 1135 C.E., according to paleoclimate studies of luminescence, stable isotopes, and other data in a stalagmite from a cave only 15 kilometers away, done by geologist James Webster of the U.S. Environmental Protection Agency and colleagues. Moyes and Awe surveyed 53 caves in central and southern Belize and found a similar pattern of large, whole jars left on ledges during the proposed dry period—evidence, they suggest, of a previously unknown drought cult. “When things get really tough, I think the Maya elaborate on what they are willing to give these rain gods,” Moyes says. “I think they are trying to give them nicer gifts.”

Other researchers say that Moyes's work offers important insights. “This is a novel way to address the issue of drought,” says Lisa Lucero, an archaeologist at the University of Illinois, Urbana-Champaign. “And the results support data from other sources, such as lake cores.”

## Clearing forests, drying climate

If drought helped drive the collapse of many Maya cities, the Maya themselves may be partly to blame: Their own practices, such as deforestation, may have changed climate and helped speed their downfall. That idea was popularized in environmental historian Jared Diamond's 2005 book Collapse.

The Maya cut trees to plant cornfields and use as fuel, both for cooking and for heating limestone to make lime for the plaster they lavished on their building projects. Evidence suggests that at least some city-states had extensively cleared forests at the time of the collapse. A 1988 pollen analysis indicated, for example, that the Maya at Copan cleared 23 square kilometers of pine forest by 800 C.E. And in a paper now in press in Estudios de Cultura Maya, archaeologists Isabel Villaseñor of University College London and James Aimers of the State University of New York, Geneseo, demonstrate that the Maya at Palenque and Calakmul substituted inferior clay for lime in their plasters just before the Classic collapse, suggesting most trees were gone.

At the SAA meeting, a team led by archaeologist Thomas Sever of the University of Alabama, Huntsville, reported the first results from a major computer simulation of climate and deforestation in the Maya lands. “There's always been the notional concept that deforestation leads to less rainfall,” says Daniel Irwin, a research scientist at NASA in Huntsville, Alabama. “A lot of people have theorized about this impact going back in time to the Maya, but this, I believe, is the first time a research team has used climate models to demonstrate it.”

Studies show that deforestation affects climate in several ways. A logged landscape absorbs more sunshine, and its new vegetation possesses shallow roots that reduce the amount of ground water returned to the atmosphere. It also offers less resistance to wind, so gusts can quickly replace humid air with dry air. To simulate deforestation, Sever and colleague Robert Oglesby, a paleoclimatologist at the University of Nebraska, Lincoln, used two climate models and ran two extreme scenarios—complete deforestation and an entirely treed landscape—to establish the boundaries of what might have happened.

Removing all trees across the Maya territory clearly changed climate for the worse. On average, says Sever, “there was a 3° to 5°[C] increase in temperature, and this led to a 20% to 30% reduction in rainfall.” Each region was hit a little differently, however. “You get a varying amount of decreases in rainfall and increases in temperature,” says team member Robert Griffin, a Ph.D. student at PSU, “so this adds a spatial pattern to it.” Of course, not all Maya lands were deforested, and the study's next phase will examine how farm production affected the extent of deforestation and degree of drought.

Although Maya archaeologists clearly have their work cut out for them as they gather new kinds of data, some researchers think the picture is growing clearer. “Now that the data [are] coming in, we can really take the idea of climate change seriously,” says UCSB's Fagan. “I don't think that anyone is saying that drought is the only cause, but it clearly is a significant factor.”

• * Heather Pringle is a contributing editor at Archaeology magazine.

13. Newsmaker Interview

# Jim Kim on Why He Took the Top Job at Dartmouth

1. Jon Cohen

An Ivy League school has broken new ground by hiring an Asian global health specialist to build its future

Jim Yong Kim, 49, a global health leader whom Dartmouth College recently selected to be its next president, specializes in surprising juxtapositions. He is fluent in English, Korean, and Spanish. He was born in Seoul but grew up in Muscatine, Iowa. In high school, he was both valedictorian and quarter-back. In 1987, Kim and fellow medical student Paul Farmer co-founded Partners in Health, now world-renowned for improving health care in Haiti and other poor countries. After completing his M.D., he earned a Ph.D. at Harvard University in medical anthropology and then launched a Partners in Health program to combat multidrug-resistant tuberculosis in Peru. The MacArthur Foundation awarded him a “genius” fellowship in 2003, and the next year he moved to Geneva, Switzerland, to run the HIV/AIDS program at the World Health Organization. Kim returned to Harvard in 2005 to direct the François-Xavier Bagnoud Center for Health and Human Rights and focus on health care delivery issues. He will take the helm at Dartmouth in July. Kim spoke with Science on 20 March; these comments are edited for clarity and brevity.

J.Y.K.:In every person's life you have to make a choice, and it's whether you continue to throw your own body at problems or step back and say my role is going to be leadership. I didn't think I was at that stage necessarily, but when Dartmouth asked whether I was interested in this job, those are thoughts that rushed into my mind. I've taken on some major problems in the last 25 years, drug-resistant tuberculosis and HIV, and those were fantastic experiences, but I felt very, very strongly that we needed to train the next generation of young people who would do this work.

Q:How did it happen?

J.Y.K.:A colleague at Harvard, Albert Mulley, was the chair of the selection committee. We sat down one day and talked about the science of health care delivery—he'd done such great work in the United States and was getting interested in global stuff. About a week later, he called me and said, “Hey, Jim, did you ever think about being a college president?” I said, “Sure, why not?” Here's why: One of my most important mentors, Howard Hiatt, former dean of the Harvard School of Public Health, said to me, “When people come to you and ask you to look at an academic job, even if you don't think that you'd take it, it's your duty as a citizen of the academic world to go look at it, think about it, and tell them what you would do if you were offered the job.” I met with them four times. At each meeting I said, “I still don't understand why you guys would be interested in me, but if you were to offer me the job, here's what I'd do.”

Q:And what's that?

J.Y.K.:One of the major crises in the world today is how we poorly execute around our most cherished social goals: health care, education, and social welfare. There's this overall ideology that once you have the pill, the surgical procedure, the content and the curriculum of a university course, the execution and delivery will take care of itself. Having the product is just the beginning. What I am most excited about in terms of going to Dartmouth is to really try and see if we can develop a science and an extraordinarily good practice in delivering around these goals. And it's about leadership, management, complex social organizations, engineering, and a lot of anthropology.

Q:Dartmouth emphasized that you're the first Asian to lead an elite Ivy League school. Does that have meaning to you?

J.Y.K.:It has a lot of meaning to my mother. And it seems to have a lot of meaning to the people in Korea. The Korean press has really responded to this with great energy. It's almost like I broke a barrier for Koreans.

Q:You're also the first global health specialist to lead an Ivy.

J.Y.K.:People who do global health now are not thought of as these fringe lunatics who wear boots and get malaria but [are thought of] as people who are relevant in the world today. We've kind of come of age if one of us can become president of an Ivy League institution.

Q:Your focus has always been on diseases of poverty. This is a major shift away from that, isn't it?

J.Y.K.:If I can train a whole bunch of people who are really committed to making the world better, and if as university president I continue to talk about social justice and health care in a way that makes sense to a broad population, I'm not sure my impact on global health is going to be diminished. I'm not going to Dartmouth to do global health; I'm going to Dartmouth to be president. But it turns out that the stuff I'm working on really fits with Dartmouth. It's a university that has a history of engaging with the rest of the world, and it's more committed to building the best possible undergraduate education than any institution I've ever seen.

14. Neuroscience

# A Quest for Compassion

1. Greg Miller

Guided by a passionate leader, a new research institute hopes to draw lessons from Buddhism to study altruism and make the world a better place

Back in 2000, James Doty was living the high life. He drove a Ferrari to work and was in the process of buying a private island in New Zealand, a villa in Tuscany, and a penthouse apartment in San Francisco. A neurosurgeon turned biotech investor, Doty had amassed more than $70 million, at least on paper. At 45, he was planning to retire, donate a large chunk of his fortune to charity, and divide his time between his three idyllic homes while doing medical volunteer work in developing countries. Last month, Doty was standing behind a lectern at Stanford University in Palo Alto, California, explaining how he'd lost it all in the dot-com bust. “Within 6 weeks, I was$3 million in the hole,” he said. He kissed the island, villa, and penthouse goodbye. But he decided, against the advice of friends and family, to follow through with stock donations that he'd promised before the crash to a handful of universities and health charities. (By holding on to the stock until the market recovered, the recipients ultimately received nearly $30 million.) Doty says that losing his material wealth made him more reflective. “Becoming completely detached from something you think you need is an interesting exercise,” he said, his voice catching with emotion. “What you realize is … it doesn't define you as a person.” His face flushed, he seemed unable to continue. Uncertainly at first, people began to clap. It was an unusually personal speech for an academic conference, but it was also an unusual conference. The audience included psychologists, philosophers, economists, neuroscientists, and theologians who'd gathered for 2 days to inaugurate a new center at Stanford for the scientific study of compassion. Doty is the co-founder and director, and he provided$150,000 to get it started. The program has also received $1 million each from two Silicon Valley investors and$150,000 from the Dalai Lama, the most he has ever contributed to a research project.

The new Center for Compassion and Altruism Research and Education (CCARE) will study the biological roots of benevolent behavior and investigate whether mental exercises derived from the centuries-old tradition of Buddhist compassion meditation—but stripped of religious trappings—can foster compassion in nonbelievers. Doty says he is not a Buddhist and does not meditate, but he thinks such exercises could find many uses. Earlier this year, for example, he flew to Washington, D.C., to talk with military leaders about treating frontline medics and chaplains suffering from “compassion fatigue”—a form of traumatic stress brought on by caring for traumatized soldiers.

Doing good science in this area is tricky business, cautions CCARE co-founder William Mobley, a neurologist and soon-to-be head of the neurosciences department at the University of California, San Diego. On one hand, Mobley says, there's the risk of researchers' personal beliefs interfering with their objectivity. “Scientists are supposed to be professional skeptics, but there are people in the field … whose only interest is seeing God's face.” And then there are experimental limitations, such as having to rely on first-person accounts of what's going through someone's mind during meditation. Even so, Mobley says, scientists who dismiss such work out of hand—and he has heard from plenty of them—are misguided. “Nothing is off-limits to science and critical thinking,” he says. “We don't have great tools, but they're good enough to get started.”

## Defining compassion

One starting point for scientific inquiry is a clear definition of the object of study. Yet the recent CCARE meeting made clear that compassion is hard to pin down. If the participants had chosen, they might have drawn up a Venn diagram of overlapping terms—sympathy, empathy, and altruism, to name a few—preferred by scholars from different disciplines. Although most participants seemed receptive, the juxtaposition of science and faith wasn't to everyone's liking. “All of this quoting His Holiness left and right makes me a little queasy,” grumbled one scientist.

Indeed, the meeting got off to a spiritual start with a haunting cello solo, The Invocation for World Peace, played by its composer, Michael Fitzpatrick, who has performed for the Dalai Lama. Thupten Jinpa, a former Buddhist monk who has been the principal English translator for the Dalai Lama since 1985, then described the long history of examining compassion in philosophy and religion and urged scientists to “approach the field with openness and perhaps humility.” Central to the Buddhist concept of compassion, Jinpa explained, is the belief that the capacity to feel—and wish to relieve—the suffering of others is inborn. The ability to feel compassion for family and loved ones comes naturally, but people must deliberately cultivate compassion for wider “circles of concern,” Jinpa said. “The highest form of compassion transcends all boundaries and embraces all sentient beings.”

Psychologist Paul Ekman, a professor emeritus at the University of California, San Francisco, whose pioneering work on facial expressions and body language was the inspiration for the new TV series Lie to Me, spoke of the “amazing coincidence” between Charles Darwin's views on compassion and morality and those of Buddhism. Ekman noted that in The Descent of Man, Darwin wrote that one of humankind's noblest virtues “seems to arise incidentally from our sympathies becoming more tender and more widely diffused, until they are extended to all sentient beings.” Ekman recounted reading the passage to the Dalai Lama (with whom he co-authored a book last year on emotion and compassion). When he was sure he had understood correctly, His Holiness declared: “I am a Darwinian,” Ekman recalled.

Ekman also noted that Darwin's observation that the sight of suffering causes pain in those who observe it meshes well with the Buddhist notion that compassion involves feeling the suffering of others as unbearable. The idea also has a parallel in the research of cognitive neuroscientist Tania Singer of the University of Zurich in Switzerland, who reviewed her studies showing that brain areas involved in pain perception also respond when people observe another person wince from an electric shock (Science, 20 February 2004, p. 1121). Singer described newer evidence that one particular area, the anterior insula, has a key role in empathy. But, she added, such work points to a neural mechanism for just one facet of compassion, which, by most definitions, requires not just recognizing suffering in others but feeling compelled to do something about it.

The latter motivation appears to be innate, said Felix Warneken of the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany. He showed videos of experiments in which children as young as 18 months spontaneously came to his aid as he pretended to struggle with various tasks—reaching for a marker he'd dropped on the floor, for instance. Even chimpanzees sometimes give unsolicited help, Warneken and colleagues have found (Science, 3 March 2006, p. 1301), suggesting that spontaneous altruism is not uniquely human.

Philosopher Owen Flanagan of Duke University in Durham, North Carolina, prompted a lively discussion by asking whether compassion might be overrated. Contemporary moral philosophers and psychologists have suggested that morality comprises different modules (Science, 18 May 2007, p. 998), and Flanagan argued that putting too much emphasis on any single module, such as compassion, at the expense of others, such as a sense of fairness and justice, may not be in a society's best interest.

## Translational meditation

Doty, whose life has cast him in the roles of both benefactor and beneficiary, is certain that more compassion would be a good thing. “I grew up in poverty and had a mother who was an invalid and a father who was an alcoholic,” Doty said in a recent interview at his office at El Camino Hospital in Palo Alto, where he is again a practicing neurosurgeon. When he was 13, a local woman befriended him and taught him to meditate, which he says gave him the confidence and sense of control he needed to achieve his considerable professional and financial success. Having attained that mental fortitude, Doty says he stopped meditating in college.

After hearing the Dalai Lama speak at Stanford in 2005 on a panel with several scientists organized by Mobley, Doty decided he wanted to foster more research in this area. After many discussions with other researchers, he and Mobley launched CCARE in January.

Several pilot studies funded by the project are getting under way, including a brain-imaging study with novice and expert meditators and a collaborative study between neuroscientists and economists that will be among the first to investigate the effects of charitable giving on its recipients—in this case undergraduate students who receive financial aid.

Stanford psychologist Jeanne Tsai recently completed a CCARE-funded study on a compassion-training protocol developed by Jinpa, who is currently a visiting scholar at the school. Undergraduate students without extensive experience with meditation or Buddhism took a weekly, 2-hour course in which they first learned meditation basics such as posture and breathing techniques. Next, over the course of 6 to 8 weeks, a trainer instructed them to picture a loved one as vividly as possible and concentrate on the sense of concern they feel for this person's well-being. In later sessions, they envisioned people they knew less well or even disliked and gradually expanded this concern to them.

Tsai randomly assigned 100 willing undergraduate students to receive the compassion training, training in mindfulness mediation, or classes in improvisational theater—to control for the possibility that simply learning a new skill or engaging in a new social activity is enough to elicit acts of kindness. (Volunteers were told they'd be participating in a study to evaluate several classes thought to improve physical and mental health.) Online questionnaires probed for changes in things such as empathic concern and the tendency to take another person's perspective. Participants also kept a daily diary of “positive and negative events,” which the researchers are now combing for evidence of an uptick in compassionate acts. At the end of the experiment, participants read a letter written by a prisoner seeking correspondents and were given an opportunity to write back and/or donate money to a program aimed at stopping abuse inside prisons. Tsai says her group is now analyzing the data to see whether people who got compassion training wrote or donated more than those who didn't.

A similar study is getting under way to investigate whether compassion training for medical students might improve their bed-side manner. If it works, it would illustrate Doty's greatest hope for CCARE: to take a centuries-old religious practice and extract from it a set of mental exercises with no religious overtones that can be scientifically proven to change the way people treat each other. It's a tall order, but without passion like Doty's, it wouldn't stand a chance.

15. Astronomy

# Stars in Dusty Filing Cabinets

1. Yudhijit Bhattacharjee

A campaign to digitize old sky photographs is squeezing new discoveries out of observations dating back to the mid-19th century.

In 1962, astronomers discovered a shining dot in the sky that appeared to be moving at an astonishing 47,000 kilometers per second, or one-sixth the speed of light. The velocity indicated that the object—named 3C 273—was a few billion light-years away, yet it was so bright it could have been a nearby star.

To study the object further, researchers delved into a trove of the astronomical past: a collection of photographic plates at Harvard University dating as far back as the 1860s. They spotted 3C 273 on some 600 photographs taken with a variety of telescopes over 70 years, some of them days apart. The images showed fluctuations in the object's brightness on time scales as short as a week. Because the object could not be dimming or brightening faster than light could traverse it, the researchers inferred that in spite of being more luminous than a billion suns, the object had to be less than a light-week across—the size of the solar system. The finding helped characterize 3C 273 as a new type of object known as a quasar, one of the most powerful energy sources in the universe.

The discovery shows the value of historical sky observations, says Harvard astronomer Jonathan Grindlay, who is leading an initiative to scan the 500,000 plates in the university's collection and put them online. The project—called Digital Access to a Sky Century at Harvard (DASCH)—is part of a movement by a small but persistent group of astronomers to preserve, digitize, and study old astronomical photographs in hope of doing new science.

Proponents argue that old plates provide the only way modern astronomers can study astrophysical phenomena on time scales longer than a few decades. “Why would you want to wait another 100 years to learn how certain stars might be varying in brightness and position over long time periods when we have this resource right here in front of us?” asks Grindlay, referring to the Harvard collection.

Preserving and scanning old plates, however, has been slow to win support from the broader astronomy community and funding agencies. Universities and observatories often discard plate collections when astronomers retire. Digitization projects in the United States and Europe—including DASCH—have proceeded in fits and starts on shoestring budgets.

“We live in a world where money is fixed—so the question is, what is the relative merit of the old data compared to new data?” says David Monet, an astronomer with the U.S. Naval Observatory's (USNO's) station in Flagstaff, Arizona, who until 2000 led the scanning of some 20,000 old plates for a searchable online sky catalog. Although he spent nearly 15 years on that project, Monet now thinks historical observations are of little value because of limitations on how accurately the brightness and position of objects can be determined on the images. “The thrill of going back 50 years” is one thing, he says, but “is the science case for doing so strong enough?”

Absolutely, say proponents, citing hundreds of newly identified variable stars in the tiny fraction of Harvard plates digitized to date. Meanwhile, the movement to digitize archives is getting a push from the Pisgah Astronomical Research Institute (PARI), a nonprofit in Rosman, North Carolina, which has started acquiring plate collections from institutions that no longer have room to house them. “Each of these collections is like a time machine. There's no substitute for having them,” says Geoffrey Clayton, an astronomer at Louisiana State University (LSU) in Baton Rouge. “Even if you can't think of what can be done with them today, it's tremendously important that they be preserved.”

## Galaxies on glass

The Harvard collection occupies parts of three floors of an old brick building at Harvard College Observatory. The plates, slabs of thin glass coated with photosensitive emulsion, sit in crowded rows of green metal cabinets. Each rests in a yellowish brown envelope marked with handwritten details such as when the photograph was taken and with what instrument. The oldest is an image of the moon from 1849. Many show galaxies—a dot with a milky swirl around it, tiny yet somehow perfect. One plate shows the brilliant tail of comet Halley streaking across the sky in 1910.

The photographs were taken with some 20 telescopes operated by Harvard astronomers at the observatory, in other parts of the United States, and at remote observing stations such as Arequipa Observatory on Monte Blanco, Peru. Around 1881, as more and more plates started coming in, the then-director of the observatory, Edward Pickering, realized that the work of documenting the positions and magnitudes of imaged objects had to be sped up. He put his Scottish maid, Williamina Fleming, on the job. She showed such talent that the observatory soon hired a legion of women, later known as the Harvard Computers, to catalog the observations.

Grindlay and his colleagues began digitizing the plates in 2008, using a souped-up commercial scanner and software specially written to translate the scanned images into crunchable data. The first step in the scanning process is cleaning each plate. “There's 100 years of Cambridge grime on them. If we didn't clean them, there would be 10,000 more stars per plate,” says Alison Doane, curator of the collection. Doane and her assistant then transfer the plates to the scanning room where Edward Los, a retired software engineer, loads two at a time onto the machine. The plates move on a cushion of air, like a puck on an air-hockey table, for precise alignment. After 90 seconds, an image of each plate appears on a nearby computer screen.

## Digital haystacks

Usually, as in the search for 3C 273, researchers turn to archival plates in search of specific objects that have grabbed their attention. Bradley Schaefer, an LSU astronomer, is typical. In recent years, he has searched the Harvard archives and the collection of 300,000 plates at the 84-year-old Sonneberg Observatory in Germany for past records of recurrent novae. Recurrent novae (RNe) are white dwarfs that erupt in brilliance every few years or decades as they capture material from an orbiting companion star.

Schaefer says he studied thousands of images of sky regions where sightings of any of the 10 known galactic RNes had been reported. The tedious task paid handsome dividends, Schaefer reported at the American Astronomical Society meeting in Long Beach, California, in January. He and colleagues, including his student Ashley Pagnotta, discovered six previously unknown eruptions, established the orbital periods for all RNes except one, and predicted when they would erupt next. Schaefer says the most significant result was detecting long-term changes in the orbital periods of two RNes, evidence that both white dwarfs were gaining mass between successive eruptions. He concludes that these two RNes will “soon enough collapse” as Type 1a supernovae—white dwarfs that burn with an extraordinarily high but fixed amount of incandescence because they've reached a critical mass from the matter acquired from the companion. He says the finding supports the idea that Type 1as are born from RNes.

Grindlay says digitizing old plates can allow researchers to find interesting phenomena instead of being limited to searching for specif ic objects. “Once you've identified your needle in the haystack, then you can figure out what's going on with it,” he says. His doctoral student, Sumin Tang, has been trying out the approach using an algorithm to search for variable stars on the 1500 or so plates that have been scanned so far at Harvard. Among three long-term variable stars she has found is one that appears to have dimmed significantly over the past 90 years. “One hypothesis is that the star is throwing off a dust shell to make itself obscured. Maybe we are seeing the star in an interesting stage in its evolution,” says Tang, who is following up with targeted spectroscopic measurements.

## Old data, new science

So far, the $4 million needed to scan the whole Harvard collection hasn't materialized. However, Grindlay is optimistic about persuading university administrators to make the funds available, not least because digitizing the plates would allow Harvard to reclaim the room the collection currently occupies. “You'll never find a more inexpensive way to clear out that much office space in Cambridge,” he points out. The project has until now been supported by$600,000 from the National Science Foundation, used primarily to build the scanner and develop software to handle and analyze images.

Digitization efforts are under way at other places as well. More than a third of the 300,000 plates at the Sonneberg Observatory have been scanned in the past 5 years. The scanned images are sitting on compact discs, says astronomer Peter Kroll, who formerly worked at the observatory and now heads a software company—4pi Systeme GmbH—that has run the institution since late 2003. The company, which has bankrolled the project with profits from its other ventures, hopes to put the image database on the Web in the coming months, Kroll says.

The Royal Observatory of Belgium (ROB) in Ukkel, which like Harvard has developed a special scanner for old photographic plates, is digitizing a collection of some 3000 images of the moons of Jupiter and Saturn. The observations, made by Dan Pascu of USNO in Washington, D.C., will help “recalculate the orbital parameters” of the moons and study how the orbits of certain moons like Io “have changed over time,” says Jean-Pierre de Cuyper, an astronomer at ROB, which is collaborating with USNO and the Paris Observatory.

For every collection being scanned, many others are gathering dust, says Michael Castelaz of PARI. In the past 5 years, the nonprofit has “rescued” half a dozen collections from institutions that could no longer house them. One example: some 10,000 plates from the Warner and Swasey Observatory at Case Western Reserve University in Cleveland, Ohio, which were sitting in shrink-wrapped cabinets in “the back of a large university storage room behind old office and classroom furniture,” Castelaz says.

Castelaz says it will take a lot of time and tedium to put the images online. But to him and other proponents of archival astronomy, the effort is entirely worthwhile. As they see it, the future of astronomy would be poorer without revisiting its past.