News this Week

Science  13 Feb 2009:
Vol. 323, Issue 5916, pp. 860

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution


    International Groups Battle Cholera in Zimbabwe

    1. Robert Koenig

    Zimbabwe's severe cholera outbreak has devolved in recent weeks into what one international health official calls “an extraordinary public health crisis.” The disease is starting to spill over into adjacent nations and threatens to become endemic in Zimbabwe. The World Health Organization (WHO) has taken the unusual step of setting up a Cholera Command and Control Centre in the capital of Harare to coordinate an array of international groups, including United Nations agencies, which are working around the clock to shore up health services, distribute medications, and treat water.

    Calling Zimbabwe's once-proud public health system “an empty shell,” WHO's Assistant Director-General for Health Security and Environment, David L. Heymann, told Science that the nation's public health infrastructure is so weak after years of neglect that WHO is concerned that other infectious diseases may emerge and that the epidemics of HIV/AIDS and drug-resistant tuberculosis (TB) may worsen.

    Since August, the cholera outbreak in Zimbabwe has killed some 3400 people and infected an estimated 68,000. Some experts predict that the case total will top 100,000, meaning that cholera would have struck nearly one in every 100 Zimbabweans. Lesser outbreaks are spreading across the borders into Zambia, South Africa, and Mozambique.

    In Zimbabwe, outside experts say, the outbreak resulted from the breakdown of potable water and sewage systems—a symptom of the country's economic chaos. The crisis worsened in part because the public health system—devastated by the loss of so many doctors and nurses who fled the country to make a living elsewhere—was severely understaffed and underfunded. Rather than dispensing the recommended oral rehydration salts (ORS), the government initially encouraged people with cholera to rehydrate themselves at home by drinking a solution of salt and sugar—an ineffective response because many could not afford the ingredients.

    Zimbabwe's health ministry has refocused its efforts, and WHO and other international groups are helping with several key activities: distributing ORS and chlorine tablets, finding funds to pay thousands of Zimbabwean health care workers, and providing better services in remote areas. “We are sending more teams into the provinces now,” says Dominique Legros, a medical epidemiologist with WHO's Disease Control in Humanitarian Emergencies program, who helped put together the cholera partnership in Zimbabwe. Although the overall case mortality rate remains an unacceptable 5%, Legros says the weekly mortality rate has begun to fall, recently dropping below 4%. The target in cholera outbreaks is to get below 1%.


    Its health system in tatters, Zimbabwe has been hit with a devastating cholera epidemic that has taken a heavy toll in remote areas.


    Vaccines have not been a factor. The only approved cholera vaccine available—the Swedish-produced Dukoral oral vaccine—must be kept cool and administered twice with at least a week's delay in between, which would be difficult in remote areas. There are two other cholera vaccines. But the newest, made in Vietnam, has not yet gained WHO approval, and an older one, developed by James B. Kaper of the University of Maryland's medical school and Myron M. Levine, director of the university's Center for Vaccine Development, went out of production in 2004. Levine says, “We are now trying to get the vaccine back into production.”

    It is highly unusual to find “cholera breaking out so widespread in a country, involving more than 85% of the districts,” says Pradip Bardhan, a cholera specialist who led a team from the Bangladesh-based International Centre for Diarrhoeal Disease Research to investigate the Zimbabwe outbreak. Heymann said it is ironic that cholera could take hold in a country that once had one of Africa's best health infrastructures, including a first-rate public health research center, the University of Zimbabwe's (UZ's) Blair Research Institute. Heymann says the center “has pretty much dried up.”

    UZ pharmacologist Chiedza Maponga agrees that there is a “shortage of health personnel, not just physicians but environmental health practitioners and nurses.” The number of doctors may have dropped to fewer than half the official level of about 2000 (Science, 4 May 2007, p. 684) in a nation of 12 million, which may also have been reduced to about 10 million due to a mass exodus.

    Epidemiologist Chris Beyrer of the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland, who recently visited Zimbabwe with a Physicians for Human Rights delegation, says he was appalled by the country's public health system and deeply concerned about the nation's HIV/AIDS and TB epidemics, as well as the cholera crisis. “One clinic we visited had not seen a physician since 2006,” Beyrer says, “and many health workers had not been paid in months.” He believes the international community should consider taking over the public health system “in a kind of receivership.”

    WHO estimates that about 120,000 people died of cholera worldwide in 2007 and that millions more were infected. These figures are much higher than the officially reported numbers, which in 2007 were 4031 deaths and 178,000 cholera cases. The current Zimbabwean outbreak is one of the most severe in recent southern African history. An outbreak in Angola in 2006–07 led to 85,000 cases, with a mortality rate of about 4%. A cholera outbreak in South Africa in 2002 registered 116,000 cases, with a mortality rate of less than 1%.

    Cholera had virtually disappeared from Zimbabwe over a decades-long period, even though Vibrio cholerae has been found in plankton living in the Zambezi River system. Although the pathogen has emerged several times since 1998 to infect people, health officials were able to control the minor outbreaks within a few weeks. But Legros says, “I am afraid this will not be the case this time and that it will remain endemic” in the population.


    Army Halts Work at Lab After Finding Untracked Material

    1. Yudhijit Bhattacharjee

    The U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID) has suspended research involving select agents and toxins after a spot inspection of the Frederick, Maryland, facility found four vials of Venezuelan equine encephalitis (VEE) that weren't in its electronic database. Some 350 researchers and technicians are affected by the stoppage, which began on 6 February and could last 3 months or longer as officials open up every freezer and refrigerator to take stock of all hazardous biomaterials at the institute.

    The lab, the largest U.S. defense facility for work on deadly pathogens, has been under intense scrutiny since the Federal Bureau of Investigation (FBI) named former USAMRIID researcher Bruce Ivins as the perpetrator of the 2001 anthrax letter attacks (Science, 8 August 2008, p. 754). Ivins committed suicide on 29 July 2008, but FBI officials have argued that the evidence against him is beyond doubt and that he carried out the mailings using anthrax stolen from a flask at USAMRIID. A special task force has spent the past few months considering what new measures are needed to improve security at USAMRIID and other Army biodefense labs.

    Last month, all lab administrators were ordered to file a Serious Incident Report if they discovered materials that did not have a corresponding record in the lab's computerized inventory. Traditionally, such materials would have been added to the inventory without any disruption in the workflow. “In the past, we would have entered the [four VEE] vials into our database and moved on,” USAMRIID deputy commander Mark Kortepeter told Science. But the new requirement prompted institute managers to “reexamine how we do inventory,” he says. Officials realized that “if we can find four vials in a spot check, it is possible there are other vials in our freezers that are not accounted for in our database.”

    In a 4 February memo to employees, USAMRIID commander John Skvorak explained that all work with select agents would be suspended until he received “certification that the full contents of each freezer and refrigerator have been evaluated and that all BSAT [biological select agents and toxins] is included in our inventory.” Some critical animal experiments will be allowed to continue.

    No-work zone.

    USAMRIID scientists are taking inventory instead of doing research.


    Kortepeter says officials expect to find materials in addition to those listed in the database because of human errors made when the lab switched from a paper to a computerized system in 2005. He says the four vials of VEE were part of a set of 20 and that 16 had been entered into the system. “It was simply an oversight,” he says.

    Greg Koblenz, a biosecurity expert at George Mason University in Fairfax, Virginia, believes that the work stoppage is justified. “Not having an accurate and complete baseline inventory makes it difficult to ensure that no material has been diverted,” he says. Koblenz says it's reassuring that high-level Army officials are trying to improve the system.

    Kortepeter says there are no plans for USAMRIID to investigate the possible loss or theft of previously undocumented materials. Other security measures at the institute would have detected any such incident, he says, even with an incomplete inventory.

    Researchers affected by the new rules are understandably unhappy, says Kortepeter, but they “realize the importance” of carrying out the inventory. Many of them will help with the task, he says.


    Journal Flinches as Article on Voice Analyzer Sparks Lawsuit Threat

    1. Adrian Cho

    A peer-reviewed journal recently yanked a long-since published paper from its Web site after the makers of a voice-analysis system—which is sold as a device to detect emotional stress and help ferret out liars—complained that the article contained inaccuracies and defamed them. The authors of the review article, two scientists from Sweden who normally study the sounds of speech, complain that the company is attempting to stifle free inquiry. The company founder counters that the paper was less a scientific analysis of his product than a personal attack. Meanwhile, 25 local governments in the United Kingdom are already using the controversial technology to try to weed out fraud among people applying for public assistance, and its use may be extended nationwide.

    Penned by phoneticians Anders Eriksson of the University of Gothenburg and Francisco Lacerda of Stockholm University, the controversial paper appeared in the December 2007 issue of The International Journal of Speech, Language and the Law with the title “Charlatanry in forensic speech science: A problem to be taken seriously.” The paper argues that there is no scientific basis for techniques that aim to determine emotional stress by analyzing the sound of a person's voice. In particular, Eriksson and Lacerda take aim at the Layered Voice Analysis (LVA) systems made by Nemesysco Ltd. of Netanya, Israel, of which they write, “The ideas on which the products are based are simply complete nonsense.”

    Amir Liberman, the founder and CEO of Nemesysco, took exception to the paper. On 3 November, his lawyers wrote to the journal's publisher, Equinox Publishing Ltd. of London, threatening to sue for defamation if the article was not retracted. Liberman says he acted not because the researchers questioned his technology but because they targeted him personally by, for example, noting in a section titled “Who is Mr. Liberman?” that he has no university degree. “The objection was not in the publication of their study results, it was in their calling us charlatans,” Liberman says.

    On 4 December, the journal removed the article from its subscription-based Web site and posted a notice that acknowledged that “Mr. Liberman and Nemesysco Limited … were not invited to comment on the content of the article prior to its publication where, in view of the content of the article, it would have been appropriate to invite them to do so.” Janet Joyce, managing director at Equinox, declined to discuss the specifics of the case but says the journal—which is published biannually, has a circulation of fewer than 500, and employs no full-time staff—simply lacks the resources to put up a legal fight. The journal has agreed to publish a rebuttal letter from Lieberman and the company, but Joyce notes that “we didn't withdraw the article. It's still in print.”

    In this corner.

    Nemesysco founder Amir Liberman (left) says phonetician Francisco Lacerda defamed him.


    In the paper, Eriksson and Lacerda cite two studies that question the scientific validity of LVA technology, which analyzes the pressure waves that make up an utterance and tallies spikes, or “thorns,” and plateaus in order to deduce the speaker's state of mind. The phoneticians also tried to reproduce the software for the LVA system using computer code in Liberman's 2003 patent as a guide. Lacerda claims that the spikes and plateaus are largely artifacts of the digitization of the sound and that the system makes arbitrary connections between the rates at which they occur and emotional states such as “high stress” or “untruthfulness.” “The measurements themselves … don't contain any meaningful information,” Lacerda asserts.

    Liberman counters that Eriksson and Lacerda used information from only one of three patents and that they never used one of Nemesysco's systems itself. “This attack is being made by people who never saw our technology, never touched the equipment,” he says. Yossi Pinkas, vice president for sales and marketing at Nemesysco, says that the company, which employs 11 people, has rigorous studies that show how the technology works, but “we don't necessarily publish everything that we have because we've been ripped off a couple of times.”

    The debate notwithstanding, the Nemesysco system is already in use. In an effort to stamp out fraud, 25 local government councils in the United Kingdom are using it in conjunction with scripted questions to evaluate phone callers applying for housing and other benefits. The national government will decide in spring 2010 whether to deploy the technology nationwide, says John Stevenson, a spokesperson for the Department of Work and Pensions. Officials use the system to decide whether further investigation is warranted, he says. “They wouldn't just say from one phone call we're stopping the benefit,” Stevenson says.

    The London Borough of Harrow has used the technology since May 2007. “In our view, it's been a great success,” says Harrow spokesperson Fergus Sheppard. “We estimate that we've saved £520,000.” From May 2007 to January this year, the system flagged 189 callers as “high risk,” and 84 eventually had their benefits reduced, he says. In addition, of 1948 callers flagged “no risk,” 474 voluntarily admitted that their original claims were inflated when they were told that they were being evaluated by the system. “They self-policed,” Sheppard says. “The system has a deterrent effect.”

    That's all it has going for it, Lacerda argues. He contends it's unethical for the government to use technology that, in his view, has no scientific basis and essentially deceives people—even if it only tricks would-be cheaters into going straight.


    Taking Stock of A Cell's Protein Production

    1. Robert F. Service
    Freeze frame.

    New technique shows just how efficiently ribosomes translate messenger RNA into proteins.


    High-school students learn that cells build themselves in a sort of waltz: Genes encoded in DNA are transcribed into fragments of messenger RNA (mRNA), which are then translated into proteins. But what exactly are they building? Microchip-based RNA arrays can show just which genes have been switched on to produce mRNAs, but not all RNAs get translated into proteins. Researchers would love to have a fast way of tallying all the proteins being synthesized at a given time and how far along each one is. Now that wish may be coming true.

    In a paper published online by Science this week (, researchers led by Jonathan Weissman, a Howard Hughes Medical Institute investigator at the University of California, San Francisco, report a technique that takes a snapshot of a cell's protein production: each protein's identity, how far along it is in synthesis, and how quickly and efficiently it's being made. The information is far better than RNA arrays at predicting how many proteins will wind up in the cell-information that could reveal which proteins are switched on during cancer and other diseases.

    “I'm excited about the work because it gets at the dynamics of gene expression,” says Rachel Green, a molecular biologist at Johns Hopkins University in Baltimore, Maryland. The new work “is at a resolution you just couldn't get” with coarser-grained older methods, Green says. In addition to offering insights into diseases, Green and others say the new tool could help developmental biologists understand the earliest stages in the development of life. “It allows you to ask a different set of questions in a global way,” says John Yates, a proteomics expert at the Scripps Research Institute in San Diego, California.

    Tracking protein synthesis has long been an inexact science. Several years ago, researchers led by Patrick Brown at Stanford School of Medicine in California improved matters by developing a technique for looking at mRNAs as the cellular machines called ribosomes translate them into proteins. The technique, known as polysomal profiling, takes advantage of the fact that multiple ribosomes often work simultaneously along one RNA strand. Researchers lyse cells in the presence of a chemical that stops protein synthesis and glues the ribosomes in place on the mRNAs. Then they sort mRNAs by how many ribosomes are stuck on each one, to estimate how many copies of the protein were being made. Finally, they identify the mRNAs present using RNA microarrays, thereby revealing the genes-and thus the proteins-they encoded.

    Polysomal profiling has its limits. For starters, ribosomes sometimes get stuck on regions of the mRNA that don't become part of the protein, or hit sites where they pause. Such regulatory switches can shut down a protein's production before it even starts. To get at these details, Weissman's team, working with yeast cells, first halted protein production as before. The researchers then used a standard enzyme to digest all of the mRNA except snippets buried and protected inside each ribosome. They collected those RNA snippets, converted them to DNA, and ran them through a high-throughput gene-sequencing machine. By comparing the resulting sequences with a database of yeast gene sequences, the team could identify which protein each ribosome was producing and where on the larger mRNA strand it had stopped in its tracks.

    The technique is like taking aerial pictures of cars on a highway, Weissman says. Whereas polysomal profiling counts just how many cars are on a stretch of road, the new technique shows exactly where each car is, making it possible to see whether traffic is moving smoothly or is piled up at an accident.

    Those details revealed some surprises, Weissman says. For starters, by looking at millions of ribosomes, the researchers found that they could track how efficiently mRNAs were translated into proteins. That's because they knew how many ribosomes were present on each message, how many proteins they wound up producing in the end, and exactly where ribosomal traffic jams were holding things up. The efficiency of protein production, it turns out, is far better than the number of copies of mRNA for predicting how many copies of a protein will be produced.

    Weissman's team could also see sharp differences in yeast cells under different conditions. For example, yeast cells that were starved for nutrients started producing proteins at mRNA sites that were different from those of normal cells-a way the protein translation apparatus helps regulate the amount of proteins inside each cell. “We can now look directly at [protein] translation and see a great deal of regulation not seen by microarrays,” Weissman says.

    The ability to compare protein translation among different cells and tissues could prove a boon to everyone from developmental biologists to diagnosticians, Weissman says. By combining the new technique with known methods for tagging ribosomes in specific tissues, researchers could look for changes in protein translation between animal models with and without breast cancer, for example. It may also prove useful for proteomics researchers by revealing new proteins produced in low abundance, a long-standing challenge for conventional techniques. Certainly, watching the skips and shuffles in biology's three-step is going to be a lot easier.


    Metabolite in Urine May Point To High-Risk Prostate Cancer

    1. Jennifer Couzin

    Nearly 30,000 men die of prostate cancer in the United States each year, but millions of others who have the disease are not even aware of it. This duality has long frustrated oncologists: They overtreat many men with indolent disease and at the same time fail to catch aggressive cases. “There's been a lot of screening, and people still die,” says Ian Thompson, a urologic oncologist at the University of Texas Health Science Center, San Antonio. The community is “desperate,” he says, for ways to identify life-threatening tumors.

    This week, a team based mainly at the University of Michigan, Ann Arbor, claims to have hit on a novel approach using metabolomics. That mouthful refers to the study of small molecules known as metabolites, chemical products present throughout the body. By screening samples of men's urine, blood, and tissues, the scientists identified 1126 metabolites, then winnowed the list down to six whose levels were higher in samples linked to localized prostate cancer and higher still in metastatic disease. “We were literally going in and measuring everything we could see,” says Christopher Beecher, a chemist at the University of Michigan and co-author of the study, led by Arul Chinnaiyan and published this week in Nature.

    The Michigan group focused on one metabolite in particular, called sarcosine, which may influence cell mobility and which, when added to cells in the lab, transformed normal prostate cells into invasive ones. They hope to devise a simple urine test to pick up the worst prostate cancers at an early stage. Some of the authors have a stake in a company, Metabolon, which is working to commercialize the findings.

    It will be years before that can happen, however. The findings need to be extended to a much larger group. This study included just 110 samples each of urine and blood matched with tissue that had been removed from patients. Sixteen had no sign of disease, 12 had localized prostate cancer, and 14 had metastatic cancer.

    Chemical clues.

    Metabolites in three classes of tissue—healthy (blue), localized cancer (yellow), and metastatic disease (red)—show distinct patterns.

    CREDIT: SREEKUMAR ET AL., NATURE 457, 912 (2009)

    In this case, sarcosine was also “not much better” at picking up aggressive disease than prostate-specific antigen, the protein now widely used in screening, says Cory Abate-Shen, a cancer biologist at Columbia University. Seventy-nine percent of metastatic samples contained high levels of sarcosine, compared with 42% in localized disease and none in healthy tissue. Still, says Abate-Shen, the strategy is unusually promising: “This hasn't been done before for any cancer.”

    Researchers have poured tremendous energy into early cancer detection and forecasting a cancer's path. Hundreds of studies have been published on the use of gene expression or protein analysis. But few tests have reached the clinic. Even those that have, such as an early-detection test for ovarian cancer called OvaSure, which picks up protein patterns, engender worries that they haven't been adequately validated.

    The Michigan work shows potential because it explores a whole new class of particles—metabolites—and because it seeks to distinguish between localized tumors and the more dangerous kind that spread. “It's a terrific way to conduct discovery,” says Thompson. But he believes the results must be validated by following many men going forward, tracking those that develop disease and comparing them with those who are healthy. And that will take years.

    Patients, Thompson believes, will have little interest in waiting. When the paper's published, “I'll get the usual 100 phone calls saying, ‘My biopsy was negative,’” he predicts. “‘Can I come in and get my sarcosine measured?’”


    From the Science Policy Blog

    The big news this week was the massive economic stimulus plan as it wended its way through the U.S. Congress. The Senate added $6.5 billion for the National Institutes of Health (NIH) but pared some of the billions the House of Representatives had approved for other science agencies, setting up a round of horse-trading as the House and Senate tried to come up with a final bill. Here's a roundup of other important developments, courtesy of Science's new policy blog, ScienceInsider.

    In European news, the blog returned this week to escalating tension between the French government and its scientists. Taking their cue from an angry Iraqi journalist, several hundred researchers hurled shoes at the Department of Higher Education and Research in Paris to protest hotly contested reforms. Meanwhile, France will be selling India two nuclear reactors, ending the Asian country's long history of nuclear ostracism from the West.

    Elsewhere in Europe, the U.K.'s Royal Society called for the creation of a National Institute of Infectious Diseases to protect the public from bird flu, mad cow, and other maladies. Europe as a whole wants to protect sharks from people. The European Commission tightened regulations on catch limits and shark finning on 5 February.

    In Japan, researchers are sounding an alarm about the burdens of noncommunicable diseases, particularly cancer, in Asia. They are working on plans to bring together labs and hospitals throughout Asia into a cancer-research network, and they are looking to the 20th Asia Pacific Cancer Conference, to be held in Tsukuba, Japan, from 12 to 14 November, to give the effort a boost.

    In the United States, Senator Charles Grassley (R-IA) has introduced an amendment that would force NIH to keep a tighter leash on its grantees. He wants anyone getting more than $250,000 in grant money to be up front about potential conflicts of interest.

    For the full postings and more, go to


    Tales of a Prehistoric Human Genome

    1. Elizabeth Pennisi*
    1. With reporting by Ann Gibbons.

    After a mad scramble, researchers have completed a rough draft of a female Neandertal genome, which will offer a new view of Homo sapiens as well as our extinct cousins.

    After a mad scramble, researchers have completed a rough draft of a female Neandertal genome, which will offer a new view of Homo sapiens as well as our extinct cousins

    Coming alive.

    Two Neandertal women, perhaps resembling the one reconstructed here, lent their ancient DNA to the genome project.


    A half a gram barely tips the postal scale. But from a Neandertal fossil, it's a whopping big chunk of priceless material. Yet Croatian geologist Ivan Gušić readily gave that up on the gamble that 38,000-year-old bones from a cave in northwestern Croatia might help provide a glimpse of the Neandertal genome. This week, researchers announced that Gušić's gamble paid off: They have gotten their first peek at 3 billion bases of Neandertal DNA, and the view, although still hazy, is spectacular. Paleogeneticist Svante Pääbo of the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, and his colleagues have compiled a very rough draft of this genome, they reported in a press conference in Leipzig and in a talk at the annual meeting of the American Association for the Advancement of Science (Science's publisher) in Chicago, Illinois, this week.

    Because Homo neanderthalensis is a member of the human family, much closer kin to us than are chimpanzees, matching Neandertal DNA against our own will reveal genetic changes that define who we are, as well as a different way to be human. Pääbo says he can't wait to finish crunching the sequence through their computers. “We will have in hand hard data on what nucleotide changes occurred in our [lineage].” Those changes are the genetic basis of what makes our species unique, or “what really made modern humans ‘modern,’” as paleoanthropologist Jean-Jacques Hublin of Max Planck puts it. Initial comparisons with our own 3 billion bases indicate that a mere 1000 to 2000 amino acid differences, as well as a yet-unknown number of non-coding changes, do that job. For comparison, about 50,000 amino acid differences separate us and chimpanzees.

    More than just another genome to add to those of the chimpanzee, macaque, and several modern humans, this ancient genome marks “a philosophical turning point in our understanding of ourselves and our evolution,” says Carles Lalueza-Fox, a paleogeneticist at the University of Barcelona in Spain. A feat that 5 years ago would have required many grams of material and many more times the $6.4 million it cost to achieve, the sequencing project “calls attention to the almost unlimited opportunities that modern sequencing methods are creating,” adds population geneticist Montgomery Slatkin of the University of California (UC), Berkeley.

    The announcement comes more than a decade after Pääbo first demonstrated it was possible to get DNA from fossils of the extinct Neandertals, who lived in Europe from at least 350,000 to about 30,000 years ago. “Only a decade ago, we were hoping to perhaps see the chimpanzee genome someday. At that time, we did not even imagine the possibility of seeing a Neandertal genome,” says Ajit Varki of UC San Diego. “It's fascinating to get the first view of the genome of our closest extinct evolutionary cousins, who were so similar and yet so different from us.”

    The genome is compiled from three shards of limb bone from Vindija Cave that turned out to be from two females. With publication and release of the data expected in the next 6 months, our ability to examine the molecular details of human evolution is poised to explode. So far, the new data suggest that the human and Neandertal lineages began to diverge some 800,000 years ago, in line with the most recent estimates from partial genomic data, Pääbo reported at the press conference. Early analyses have yielded no sign of introgression of modern genes into the Neandertal sequence, supporting the idea that Neandertals did not interbreed with modern humans during the thousands of years the two species shared territory in Europe (see p. 870).

    For now, there's not enough sequence to do more than make a rough sketch of Neandertals. “Many gaps and errors are expected,” says Lalueza-Fox. To be certain about the order of the bases, sequencers need to determine each base at each location multiple times—and Pääbo has only enough sequence to cover the genome one time over. In total, the team has sequenced 3 billion bases, not quite the length of the whole genome. But some bases were sequenced repeatedly, and some spots were missed entirely. All together, Pääbo estimates they have covered 60% of the entire genome.

    “There are substantial limitations to the utility of a low-coverage genome,” says Chris Ponting, a genomicist at the University of Oxford in the United Kingdom. “[It] provides only tantalizing research leads that will always need to be validated by more in-depth sequencing, often from other fossils.” That's why other groups are seeking Neandertal DNA for their own projects (see p. 868). There are also lingering concerns about contamination by modern human DNA, which can easily masquerade as Neandertal material because the two genomes are so similar.

    Pääbo says he's just getting started. “1× is really just a milestone,” he notes. “Our ambition is to go on and produce a Neandertal genome of a quality comparable to, for example, the chimpanzee genome, over the next few years.”


    Inching toward a genome

    Ancient DNA has had a checkered past. Once organisms die, their genetic material begins to degrade. Chromosomes splinter, and some bases transform. Microbes infiltrate the decaying organism, leaving their own genomes to posterity. Humans collecting and cleaning fossils add their own DNA smudges.

    Neandertal in hand.

    Pääbo's dogged pursuit of this ancient human's DNA has paid off.


    These molecular booby traps have foiled many a sequencing effort, yielding sequences consisting mainly of contaminant. But over the past 20 years, Pääbo and others have worked to avoid the traps. In 1997, he and student Matthias Krings, then at the University of Munich in Germany, decoded about 400 bases from Neandertal mitochondrial DNA (mtDNA) (Science, 11 July 1997, p. 176). These were different enough from H. sapiens mtDNA to confirm for many that Neandertals were a separate species from us. That feat was possible because mitochondria are plentiful in cells, providing a relative mother lode of genetic material, some 500 to 1000 times more than the DNA in the nucleus.

    But in the years following, the field of ancient DNA was at an impasse. Much-heralded fossil bee DNA encased in amber and even dinosaur DNA turned out to be modern human contamination. It was almost impossible to reliably isolate enough ancient genetic material to work with. Not until 8 years later did new technologies help ancient DNA break free of these limitations.

    In 2005, James Noonan and Edward Rubin of the U.S. Department of Energy Joint Genome Institute in Walnut Creek, California, working with Pääbo, managed to extract 27,000 bases of ancient cave bear DNA using an approach originally developed for sequencing all microbial DNA in an environment (Science, 3 June 2005, p. 1401; 22 July 2005, p. 597). Six months later, Hendrik Poinar of McMaster University in Hamilton, Canada, and his colleagues used so-called next-generation sequencing technology to generate a whopping 28 million bases of DNA from a mammoth bone. “The paper opened the possibility of doing a Neandertal genomics [project] as well,” says ancient DNA expert Eske Willerslev of the University of Copenhagen. (Science, 20 January 2006, p. 392.)

    Pääbo was more than ready to put these advances to use. He and his colleagues had already collected 70 samples from bones from throughout Europe, Central Asia, and Southern Siberia. Three rather nondescript shards, collected around 1980 from the Vindija Cave in Croatia, passed initial tests for the presence of true ancient human DNA. One had a whopping 4% Neandertal DNA, compared with less than 1% in most samples tested. At 38,000 years old, these are some of the youngest Neandertal bones known, and they were buried in a cool, dry cave. And anatomically, they were uninformative, so they were unlikely to have been handled extensively.

    Priceless bone.

    This fossil fragment and two others provided the DNA for the Neandertal genome.


    Pääbo shared some of the DNA from these bones with Rubin, and the two groups sequenced it independently, using different approaches. Both succeeded: Rubin's team extracted 65,000 bases, Pääbo, 1 million bases (Science, 17 November 2006, p. 1113). Their successes prompted Pääbo, the German government, and the sequencing company 454 Life Sciences in Branford, Connecticut, to team up to do the entire Neandertal genome at 1× coverage, while the collaboration between Rubin and Pääbo ended.

    But a few observers noted that the two groups, working from the same Neandertal bone, reported different results. “The first time we looked at the two data sets, it was quite clear to us that there were large discrepancies,” Rubin recalls. For example, his group had concluded that there was no sign of modern human DNA infiltrating the Neandertal genome, whereas Pääbo's data at that time suggested that our ancestors had intermingled with Neandertals. This and other discrepancies raised suspicions that the Leipzig group's sample may have included DNA from living people. But Pääbo was convinced that contamination wasn't a problem, and given the prestige of his group, few were willing to speak out publicly at the time.

    Eleven months later, Jeffrey Wall and Sung Kim of UC San Francisco sounded an alarm. They had reanalyzed both groups' data and recalculated the divergence date of the two species. Rubin's data suggested a final split at about 325,000 years ago, which roughly concurs with fossil evidence, but Pääbo's results gave a date of only 35,000 years ago. Wall and Kim also noticed that the sequenced fragments Pääbo's group reported were relatively long, as might be expected from modern rather than ancient DNA (

    Pääbo was also finding problems with the early results. His team changed its opinion on interbreeding in May 2007 and eventually realized that 11% of their sample was modern human DNA, as they reported in the 8 August 2008 issue of Cell. “Contamination was indeed an issue,” Pääbo says.

    Microbial and nonhuman sequences among the deciphered DNA were easy to spot and cull thanks to comparative genomics, but telling Neandertal from modern human contamination was next to impossible.

    So Pääbo and colleagues, including postdoc Richard “Ed” Green, who has spearheaded the Neandertal effort, continued to improve their procedures. They did more of the sample preparation in a clean room to minimize exposure to modern human DNA. They also began adding 4-base tags to DNA extracted from the Neandertal samples so that they could detect modern human contamination introduced after the sample left the clean room. Any completed sequence lacking these tags was discarded. Says Beth Shapiro of Pennsylvania State University (PSU), University Park, “I think this tagging technique is one of the most exciting advances in using ancient DNA and provides a big step forward for the field.”

    These and other technical challenges put the team behind schedule, so they bought sequencing machines made by Illumina, a different kind of “next generation” technology that rapidly and cheaply churns out small stretches of sequence (Science, 7 November 2008, p. 838). About one-third of the sequencing was done at the 454 Life Sciences facility in Connecticut, and much of the rest rolled off the Illumina machines.

    Side by side.

    The genome will help clarify similarities and differences between Neandertals (left) and modern humans.


    Ancient panorama

    Pääbo's crew harvested one of the first fruits of the Neandertal genome this past summer, deciphering all 16,500 bases making up Neandertal mtDNA. About 200 of those bases differ from those of modern humans, supporting the idea that the two lineages were separate species, they reported in the 8 August issue of Cell.

    The mtDNA genome, plentiful in cells, was low-hanging fruit compared with the nuclear genome. It took 68.9 billion bases, 96% of which were microbial in origin, to cull the 3 billion human bases of the 1× Neandertal genome.

    One of the first results was a surprise: The indeterminate shards of bone turned out to be from two females. So this first genome carries no information on Neandertals'Y chromosome, although the group is now deciphering DNA from additional samples that may include a male.

    Now the team is matching up their 3 billion bases to those of the modern human genome to compare the two. “We hope to identify regions where the Neandertal is ancestral at positions where current humans vary,” Pääbo explains. “These will then be candidates for having undergone positive selection in recent human history.”

    The team's preliminary analysis has pinpointed many thousands of such differences. But they have yet to prove that the differences don't vary among living people. Nor have they been able to discern any patterns in the types of regions, or the genes themselves, affected by these changes.

    Until now, paleogeneticists have had to take what's called a candidate gene approach: They first identify a gene of interest and then attempt to fish it out of Neandertal DNA. In the past 18 months, those efforts have yielded intriguing insights. For example, the University of Barcelona's Lalueza-Fox found a unique Neandertal variant of a pigmentation gene, mc1r, which can result in pale skin and red hair (Science, 30 November 2007, p. 1453). More recently, Lalueza-Fox found that leg bone flakes from two adult males from the El Sidron cave in Spain had type O blood, as he and his colleagues reported 24 December 2008 in BMC Evolutionary Biology. And with Pääbo, they found that the FOXP2 gene, which has been linked to language ability, is the same in modern humans and in Neandertals (Science, 26 October 2007, p. 546.) Subsequent work suggested that the FOXP2 finding might be due to contamination, but the new genome indicates that the Croatian Neandertals had the same modern human FOXP2 variant, bolstering the notion that Neandertals may have had some linguistic facility.

    With the whole genome data now in hand, researchers can scan the entire genetic landscape for species-defining variants. The whole-genome approach “makes no assumptions about what sequences are interesting,” says Rubin. The resulting catalog of differences will be a great resource. “I don't think there is a human geneticist that has a gene or noncoding region of interest who is not going to be interested in comparing their human sequences with that of the Neandertal,” Rubin adds.

    Varki has already asked Pääbo to keep an eye out for genes for proteins that recognize sialic acids, which in chimp-human comparisons seem to be an evolutionary “hot spot,” although no one knows how the changes affect the immune and inflammatory functions of the proteins. Meanwhile, Noonan, now at Yale University, is looking at a particular stretch of regulatory DNA that has a variant unique to humans. If he substitutes the human version of this so-called enhancer for the mouse version, the mouse hand develops as if it's trying to grow a thumb. “We'd like to know when did these sequence changes arise,” says Noonan. If Neandertals—who have thumbs—don't have the modern human sequence, for example, then it's unlikely that the enhancer played a role in thumb evolution.

    Did they or didn't they?

    Everyone is eager to learn more about the perennial question of whether Neandertals and modern humans interbred. So far, early analysis of the complete nuclear genome shows no sign of admixture, says Pääbo. Working with David Reich of Harvard University and Jim Mullikin from the National Human Genome Research Institute in Rockville, Maryland, the team has compared about one-third of the Neandertal data with sequence from an African and from a European. If Neandertals and ancient Europeans had mixed genes, their genomes should resemble each other more than African and Neandertal genomes do. Thus far, the team has found no contributions by Neandertals to modern humans.

    Nor do the sequences of two genes whose histories suggested interbreeding support admixture. Researchers have suggested that a supposedly 1-million-year-old variant of Microcephalin, which is associated with brain growth and appeared in modern humans only 37,000 years ago, evolved first in the big-brained Neandertals, then spread to modern humans through interbreeding. But Pääbo's team did not find that variant in the Neandertal nuclear sequence. A gene called MAPT also has a very old variant, found primarily in Europeans, that some think may have originally come from Neandertals via interbreeding. Although the variant's function is uncertain, it today provides some female Icelanders with a reproductive advantage, and so researchers speculated that selection might have favored its spread into modern humans. But the Neandertal nuclear genome didn't show this MAPT variant, either.

    For some, the question of mixing genes was all but settled by the complete mitochondrial genome, which showed no sign of it. Still, some researchers point out that mtDNA sequence comparisons may not catch low levels of ancient interbreeding. “The problem with mtDNA is that … it offers just a single glimpse into the gene-flow question. An absence of gene flow at the mtDNA does not rule out significant levels of gene flow at other genes,” says population geneticist Jody Hey of Rutgers University in Piscataway, New Jersey. And the mitochondrial genome is inherited through only the maternal line and therefore would not preserve any sign of mixing between modern females and Neandertal males.

    With the full nuclear genome in hand—albeit from females only—population geneticist Daniel Garrigan of the University of Rochester in New York state thinks researchers should be able to detect low levels of gene flow between Neandertals and moderns. He recently used a mathematical model to test just that question. When he looked for evidence of admixture in a stretch of 1 million bases of simulated DNA code, he found that he could detect levels of admixture as low as 1% if the two species split from a common ancestor more than 600,000 years ago. “Even if admixture is low, you still have good power to detect it,” says Garrigan.

    Hands off.

    Clean-room procedures for preparing fossil samples (inset) helped reduce contamination by modern human DNA.


    In practice, however, even if Pääbo's team did see signs of mixing, they would have to be on guard against contamination, which could mimic the signature of ancient interbreeding by putting a modern human variant into the Neandertal genome. Pääbo recognizes the challenge: “It will never be possible to completely rule out minute amounts of contribution from them to us, but it becomes less and less biologically relevant the smaller that contribution could have been.”

    Reality check

    Indeed, even the most ardent Neandertal gene hunters recognize they need to temper their enthusiasm. “It will be dangerous to interpret any of the data when the coverage is 1×,” says Steve Scherer, a geneticist at the Hospital for Sick Children in Toronto, Canada. It takes sequencing a base multiple times at each spot to guarantee accuracy—and 1× doesn't allow for much redundancy. With ancient DNA, there's always a chance that the sequence is wrong because DNA changes as it ages. “Discerning what those [genetic] changes mean or meant is going to be tricky,” warns anthropologist Pat Shipman of PSU. “There are bound to be dead-ends and false leads.” In addition, this scant coverage will make finding duplicated genes and other so-called structural changes, many of which are proving important in separating mammalian species, next to impossible.

    Contamination remains a concern. “Although the criteria have improved, it still remains one of the biggest challenges to the Neandertal genome project,” says the University of Copenhagen's Willerslev.

    Also, the best way to understand a difference between, say, an H. sapiens and a Neandertal gene is to look at that base in multiple individuals. We can do that for our species, but to date, this Neandertal genome represents a consensus; individual variation and sequence validity are not yet established for many regions of the genome. It will be especially hard to trust any Neandertal base that is different from both the modern human's and the chimp's, Pääbo points out because the difference could be due to degradation.

    Much legwork needs to be done to begin to pin down genetic differences that might explain the Neandertal's bigger nose, wider body, and short limbs, as few gene variants have been tied to these traits. Lalueza-Fox is beginning to look at skeletal variation in human populations with an eye toward eventually identifying the gene variants that underlie them. Then he plans to look at those same genes in the Neandertal. “But the first step is to do the work in humans,” he says.

    Moreover, there are limits to what genes can say about how Neandertals lived their lives. “Issues of behavior and ecology will not have strictly genetic answers,” says Shipman. “In the end, it is the fossils and the artifacts that are the final arbiters about what the genes produced or did not produce.”

    But for now, the genes have turned a few hundred milligrams of bone into a gold mine. The Neandertal women who left their bones in the cave likely led undistinguished lives, yet now “they are the genetic spokespeople of their kind,” says Green. “Through the DNA preserved in their bones, they may wind up making a profound, posthumous contribution to our species' understanding of its own evolutionary history.”


    Wanted: Clean Neandertal DNA

    1. Elizabeth Pennisi

    Now that one group has finished a rough draft of the genome, other groups are working to remain in the Neandertal sequencing game—but first they need to find well-preserved Neandertal DNA.

    It usually pays to have more than one team working on a scientific problem. Take the early work on sequencing Neandertal nuclear DNA, for example. In 2005, Svante Pääbo and his crew at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, gave Neandertal DNA samples to Edward Rubin's group at the U.S. Department of Energy Joint Genome Institute (JGI) in Walnut Creek, California. Together, the two groups helped revitalize the field of ancient DNA by producing more than a million bases of Neandertal nuclear genome sequence (Science, 17 November 2006, p. 1068). Equally important: Differences in their results alerted the field to what later proved to be contaminating DNA in Pääbo's analysis.

    Siberian search.

    Edward Rubin (right) and Institute of Archaeology and Ethnography Director Anatoly Derevianko are still seeking DNA-rich Neandertals.


    But since 2006, when the company 454 Life Sciences, with its faster, cheaper sequencing technology, stepped up to do the entire genome with Pääbo (see main text), the Leipzig group has been pretty much on its own. Rubin and others are hoping, however, to remain in the game—if they can get their hands on enough Neandertal DNA.

    Rubin plans to sequence target regions from several different Neandertals to complement Pääbo's approach of massively sequencing one sample. The goal is to determine whether differences noted between Neandertal and Homo sapiens genomes are real or simply reflect variation among individual Neandertals. Alan Cooper of the University of Adelaide in Australia is also working toward scanning DNA from several Neandertals, and Pääbo himself has sequenced a million bases from additional samples.

    But Neandertal DNA preserved and excavated under the right conditions is proving elusive. Over the past 2 years, Rubin and his postdoctoral fellows have traveled the world, striking up relationships with paleontologists in hopes of finding relatively uncontaminated, undegraded Neandertal fossils. So far, they have had little luck with 20 Neandertal specimens from Israel, Spain, France, or Siberia. “None of these have had significant fractions of uncontaminated Neandertal DNA,” Rubin laments. “As it turns out, the sequencing of a Neandertal is now much easier than finding a good uncontaminated [sample].”

    In the meantime, Rubin's team has been developing methods to extract ever more DNA from each sample. Rubin hopes they will soon get to try these approaches on Neandertals. “We do have some other Neandertal specimens coming in that we are optimistic about,” he says. Ancient DNA expert Eske Willerslev of the University of Copenhagen agrees that it's important to have more than one group producing data. “It's definitely worthwhile to have independent replication of results,” he says.


    A Neandertal Primer

    1. Michael Balter

    The rough draft of the Neandertal nuclear genome may usher in a brave new world of research on these extinct humans, but after 150 years of study, we already know a few things about them.

    Why are they called Neandertals? The first Neandertal fossil was found in 1830 in Belgium, but scholars failed to realize it was an extinct human. Then in 1848, when quarry workers in Gibraltar came across a strange-looking skull, local scientists again did not know what to make of it. Only in 1856, after miners working in Germany's Neander Valley (Neander Tal in German) uncovered a skull cap and other bones, did scientists figure out what they had.

    When and where did they live? Bones of full-fledged Neandertals show up in the fossil record at least 130,000 years ago, from Spain to Uzbekistan (see map). Recently, Neandertal mitochondrial DNA was found in bones in southern Siberia, extending their apparent range another 2000 kilometers east. The last known Neandertals survived until perhaps 28,000 years ago on Gibraltar.

    At home in Eurasia.

    Neandertals ranged from Europe to southern Siberia.


    How are they related to modern humans? Neandertals are our closest relatives. Although some researchers once thought they were our immediate ancestors in Europe, most now agree that Neandertals and modern humans shared a common ancestor that lived somewhat less than 500,000 years ago, possibly in Africa. No later than 350,000 years ago, fossils resembling this ancestor but with incipient Neandertal features show up in Spain and Britain.

    Did they have sex with modern humans? If they did, the two species apparently didn't have many babies together. The latest data from the Neandertal mitochondrial genome and preliminary results from the nuclear genome show little sign of interbreeding (see main text). But many researchers say some admixture cannot be ruled out.

    What did they look like? Neandertals were once portrayed as brutish creatures, but scientists now think they resembled modern humans in many ways. Indeed, the late anthropologist Carleton Coon once suggested that a Neandertal dressed in a suit and hat riding the New York City subway would go unnoticed. The next time you ride the subway, look for someone with a stocky, muscular body with short forearms and legs; a large head with bony brow ridges; a jutting face with a very big nose; and perhaps reddish hair and fair skin.

    What did they wear? Clothing isn't preserved in the archaeological record, but researchers assume that Neandertals could not have survived ice-age winters without covering up with animal skins. Neandertals did not have needles, but some researchers have argued that the bone awls they routinely made were used to tie animal skins together.

    What did they eat? Small and medium-sized game, mostly rabbits and deer, which they probably chased and killed with spears. Some Neandertals ate seafood, including shellfish and sea birds. And at least on occasion, they may have eaten each other. At Neandertal sites in France, Spain, and Croatia, researchers have found Neandertal bones that were apparently cut with stone tools.

    Were they smart? Neandertals had big brains, ranging from 1200 to 1600 cubic centimeters, slightly larger than the modern human average. Yet judging from the artifacts they left behind, many researchers have assumed that Neandertals were not as intelligent as ancient Homo sapiens. Until the last 10,000 or so years of their existence, Neandertals exhibited little evidence of symbolic behavior, such as art or personal ornamentation. Their skillfully made tools were more sophisticated than those of their ancestors, but those tools changed little for 100,000 years. And although there is good evidence that Neandertals intentionally buried their dead, researchers argue about whether there are signs of burial rituals. Only after 45,000 years ago, when modern humans began colonizing Eurasia, do beads and more sophisticated bone tools crop up at a small number of Neandertal sites. That has led many researchers to suggest that they were merely imitating the modern humans, although that idea is bitterly debated. As anthropologist Leslie Aiello, president of the Wenner-Gren Foundation for Anthropological Research in New York City, put it: “The Neandertals had big brains, and they must have been using them for something.”

    Could they talk? They may have spoken, but probably not as precisely or inventively as modern humans do. The soft tissues most critical to human language, such as the larynx and vocal cords, do not fossilize, and there are few differences between Neandertals and modern humans in their underlying bony structures. Yet some researchers have suggested that Neandertals couldn't produce vowel sounds as well as modern humans do, and others insist that their limited symbolic behavior suggests limited language skills as well. Nor have genetic studies been able to resolve the debate.

    Why did they go extinct? Several decades ago, many researchers assumed that Neandertals met a violent end at the hands of invading modern humans. That view is no longer widespread: There is no evidence of warfare between the species, and Neandertals held on for some 15,000 years after H. sapiens arrived. Today's debates focus primarily on whether modern humans outcompeted Neandertals for resources or whether Neandertals succumbed to the cold as the last ice age approached its peak about 25,000 years ago.


    Phoenix Rose Again, But Not All Worked Out as Planned

    1. Richard A. Kerr

    For better and for worse, Phoenix often wandered from its scripted mission on Mars, but there was some groundbreaking science behind the often distracting headlines.

    For better and for worse, Phoenix often wandered from its scripted mission on Mars, but there was some groundbreaking science behind the often distracting headlines

    At first, it was all about asparagus. Early results from the Phoenix lander released at a NASA press conference showed that the soil there would be great for raising the vegetable, if only the temperature were 100°C or so higher. Then the news was the umpteenth discovery of water on Mars, even though researchers running an orbiting spectrometer had confidently predicted Phoenix's discovery of near-surface ice. And later, the big science news was snow. A laser shot skyward had detected the white stuff overhead, although it never made it as far as the ground.

    Diggin' it.

    The Phoenix lander handily scooped up soil (foreground) to reveal ice (inset), but the history of water proved more elusive.


    Throw in harrowing accounts of balky equipment, frustratingly sticky soil samples, and terminal hypothermia, and it was hard to remember why NASA sent Phoenix to Mars in the first place: to try to decipher the history of water and to see whether, at some point in that history, liquid water could have let martian life bloom. The $450 million mission of exploration fell well short of those ambitious goals, largely because Mars failed to cooperate. Phoenix found minerals that formed when rock interacted with liquid water, but no one knows for sure where or when that happened.

    “If we were looking for answers, we'd be disappointed,” says applied physicist and instrument lead Michael Hecht of NASA's Jet Propulsion Laboratory (JPL) in Pasadena, California. Even so, planetary scientists see Phoenix's open-ended exploration as well worthwhile. “It reveals that Mars is complex and heterogeneous and has a far more interesting history than thought,” says soil scientist Ronald Amundson of the University of California, Berkeley, who is not on the Phoenix team.

    Ever onward

    Before it could uncover intriguing complexities, Phoenix had to surmount some daunting problems, some of them self-inflicted. According to the plan, said Phoenix project manager Barry Goldstein of JPL, team members' “7 minutes of terror” between Phoenix blazing into the atmosphere and gently settling onto the surface was “going to be followed by 3 months of joy.”

    Not quite. Mission engineers and scientists had only 151 martian days between landing and being snuffed out by the frigid grip of approaching martian winter, and they spent them grappling with a series of time-consuming mechanical and environmental problems. Clumpy polar soil got stuck when they dumped it onto millimeter-size screening that covered instrument inlets (Science, 8 August 2008, p. 758). Sprinkling the soil worked better than dumping it, but no one is sure what caused its maddening stickiness.

    Going after ice or even icy soil—a prime target for the mission—proved nearly disastrous. Icy samples refused to fall out of an overturned scoop, like ice cubes refreezing in a drink glass. Unsuccessful efforts to work around the ice problem consumed 3 weeks mid-mission. And then there were the inlet doors of the Thermal and Evolved Gas Analyzer (TEGA), one of Phoenix's two principal analytical instruments. They wouldn't open, at least not fully. The TEGA team had caught the problem early but missed several opportunities to ensure that it had been corrected, according to TEGA lead William Boynton of the University of Arizona (AU), Tucson.

    In the end, Phoenix operators managed to fill only five of TEGA's eight single-use analysis chambers, none with the desired amount of ice. One of the four chambers of the Microscopy, Electrochemistry, and Conductivity Analyzer remained empty (though it proved to be a useful calibration blank), and MECA examined eight samples out of a possible 10 under its optical microscope. “Clearly, we would have liked to move more quickly,” says Phoenix principal investigator Peter Smith of UA. The icy soil, he says, “was just nasty. We lost part of the clues that would make interpretation easier.” Nevertheless, he adds, “we've got a lot of pieces of the puzzle.”

    A habitable world

    Many of the pieces to the water puzzle that Phoenix sent back fit surprisingly well into a picture of a habitable planet. The mission began with the assumption that the most likely place on Mars to find signs of recently liquid water—and therefore a place where microbes could have thrived—was a few centimeters beneath the surface of the northern polar region. Ice was almost certainly there, and it may have periodically melted a bit. On time scales of tens of thousands of years, the planet tilts farther over on its axis, bringing a warming sun higher in the sky of the martian arctic. That might have melted the very top of the ice, wetted the soil, altered its minerals, and—not incidentally—revived any microbes in the martian deep-freeze.

    Scientists believed that evidence of such habitable conditions would be preserved in the chemistry, mineralogy, and small-scale geology of the landing site. TEGA, they hoped, would sniff out minerals by progressively heating samples and identifying the volatile products. MECA would add its own water to the soil and see what went into solution, just as if the ice had melted.

    Phoenix certainly delivered on the habitability question. Nothing it found was inimical to life. The moderate pH of 8.3 found by MECA and the tiny amount of salts detected by both TEGA and MECA bode well for the evolution of martian life, not just for growing asparagus, notes paleontologist Andrew Knoll of Harvard University. And, much to everyone's surprise, MECA detected relatively abundant perchlorate, possibly magnesium perchlorate. Although used to fuel rockets, perchlorate can also fuel life; certain bacteria are known to use perchlorate as an energy source. “If you warmed that site up, it would be entirely habitable for some types of organisms,” says astrobiologist Dawn Sumner of the University of California, Davis.

    Phoenix also found signs of once-liquid water, although the news media paid more attention to the abortive snowstorm. MECA found at least 4% calcium carbonate and, less definitively, at least 1% sulfate, says MECA team member Samuel Kounaves of Tufts University in Medford, Massachusetts. Water plus rock typically produces those compounds. “These salts were dissolved in the past,” Kounaves says. “They have been wet before.”

    Sticky situation.

    Martian arctic soil proved to be sticky, clumpy stuff (top, in scoop) that passed through inlet screens—and partially opened doors—only with difficulty.


    Missing pages

    As promising as the habitability front looked, the other mission goal—getting a history of water on Mars—was looking tougher. “It's quite tricky,” says Smith. To start with, Phoenix dug through only about 5 centimeters of loose soil before it hit concrete-hard ice. “I don't know that we'll have much to say about profiles in 2 inches of soil,” Smith said at the meeting of the Division for Planetary Sciences last October. “That's challenging.”

    Team members had hoped to read a story about water in that soil layer. In one possible scenario, the ice would have melted tens of thousands of years ago. The meltwater would have wicked upward, leaving a thinner and thinner trail of salts behind it. Such a gradient in salt content or other soil properties from bottom to top would tell the hoped-for story. Using Phoenix's camera and the analytical instruments, “we looked for gradients,” Hecht said at the fall meeting of the American Geophysical Union last December. “We didn't see any.”

    Researchers can offer various explanations for the uninformative uniformity of the Phoenix site. “I am not surprised,” says Mars researcher John Mustard of Brown University. “You want to go down several meters,” not centimeters. Mustard sees the subsurface ice as a thick remnant of glacial ice deposited during a past martian ice age rather than a thin, soil-cementing frost formed in recent millennia. If that's true, he says, the story of water would lie far deeper than Phoenix reached.

    It's also possible that any record the thin soil held may have been wiped out. Planetary scientist Michael Mellon of both the University of Colorado, Boulder, and the Phoenix team notes that, as expected, the icy terrain around Phoenix shows signs of ever so slowly churning like a pot of boiling water as the ice's temperature swings up and down with the seasons. Such “cryoturbation” could be running fast enough to smear out any record of melting until it is unrecognizable, Mellon says. Alternatively, the wind may be to blame. Under the microscope, much of the soil appears to be either fine, windblown dust or larger, angular mineral particles that the wind bounced across the surface. That means the water-altered minerals “could come from the other side of the planet,” says geochemist Nicholas Tosca of Harvard.

    Mixed results or not, planetary scientists are more than content with Phoenix. “There are chemical processes at work we're just beginning to piece together,” says Mustard. “It's a first step.” But that first step did yield a pleasant surprise. “The demonstration of diverse chemical processes is really important for people studying Mars,” says Sumner. Observations from five previous landers and rovers at lower latitudes had painted a picture of an entire planet drenched billions of years ago in acid and brine (Science, 5 January 2007, p. 37). That was not particularly inviting for life and perhaps prohibitive for the origin of life. The Phoenix polar site, on the other hand, with its moderate pH and low saltiness, “says a huge amount in a positive way about habitability,” says Sumner. Now planetary scientists can see that, like Earth, Mars has a diversity of environments and therefore chemical and physical transitions between environments that help maintain a healthy biosphere. That, Sumner adds, “is a lot of science for the buck.”


    Can Mathematics Map the Way Toward Less-Bizarre Elections?

    1. Barry Cipra

    At the Joint Mathematics Meetings, speakers discussed ideas ranging from pie-in-the-sky theoretical to crust-on-the-ground practical for dealing with the perennially fraught issue of deciding which voters get to reelect which members of the various legislative bodies.


    With the 2010 census looming, U.S. politicians and their legal teams are gearing up for another round of wrangling over the spoils of redistricting: the process of deciding which voters get to reelect which members of the House of Representatives and other legislative bodies. Parties in power like to carve up voters to their own advantage, a practice known as gerrymandering. Some reformers, however, hope to limit the mischief—and are turning to mathematics for tools to do so. In a marathon 6-hour session at the Joint Meetings, speakers discussed ideas ranging from pie-in-the-sky theoretical to crust-on-the-ground practical.

    The term “gerrymandering” dates back to 1812, when Massachusetts Governor Elbridge Gerry signed into law a tortuous districting map that favored his Democratic- Republican Party over the rival Federalists. But given the fine-grain demographic detail of modern political databases, “the problem is much worse than it used to be,” says Richard Pildes, an expert on election law at the New York University School of Law in New York City. Gerrymandering “gives people the sense that they're not really in control of their democracy,” Pildes says. “It's part of what contributes to an alienation and cynicism about democracy.”

    The mathematics of redistricting starts with arithmetic and geometry. Ideally, every district in a state would have an equal population and would be, in some sense, both “contiguous” and “compact.” Socioeconomic, political, and racial demographics also come into play. “You can have equipopulous districts and still have whoppingly biased gerrymanders,” notes Sam Hirsch, a lawyer at Jenner & Block in Washington, D.C., who specializes in election law and voting rights.

    To a mathematician, contiguous means connected—i.e., you can travel from any point in it to any other without leaving the region. Compactness is trickier. Various definitions have been proposed, including one presented at the session by Alan Miller, a graduate student in social science at the California Institute of Technology (Caltech) in Pasadena, California.

    Miller's method, developed with Caltech economist Christopher Chambers, quantifies the “bizarreness” of geometric shapes. (The word “bizarre” traces to a 1993 ruling in which the U.S. Supreme Court struck down several oddly shaped congressional districts. Politicians' attempts to handpick their constituents invariably create convolutions in district lines.) In essence, bizarreness is the probability that the most direct path between two randomly chosen voters within a district crosses district lines. The higher the probability, the more bizarre the district is. (The path is required to stay within the state, to avoid penalizing districts that sit on ragged state boundaries.)

    Using block data from the 2000 census, Miller and Chambers have computed bizarreness for the congressional districts of Connecticut, Maryland, and New Hampshire. Most compact was Connecticut's 4th District, with bizarreness 0.023; most oddly shaped: Maryland's 3rd district, at 0.860 (see figure).

    How bizarre.

    Researchers can rank the shapes of Congressional districts ranging from highly compact (top) to convoluted.


    Bizarreness could be used as a threshold criterion in producing redistricting maps or comparing alternatives, Miller says. “You can use it to reject districts that are badly shaped.”

    In his own proposal, Hirsch took the idea of thresholds and added a dose of high-octane competition. Rival factions—or anyone else interested in entering the fray—would be able to counter one another's maps, as long as each new submission improved on at least one of three criteria and matched the other two. The goals of the three criteria are to minimize the number of counties cut up by district lines, equalize as much as possible the number of districts leaning toward each of the two major parties, and maximize the number of “competitive” districts, in which neither major-party candidate in a recent statewide contest would have won by more than 7% of the vote.

    Hirsch's proposal “is a great idea,” says Charles Hampton, a mathematician at the College of Wooster in Ohio, who has been involved in redistricting since the early 1980s. (He drew maps in 1991 for the governor of California's Independent Redistricting Panel.) “We quibble on some of the details,” Hampton says, but “I think [it] has some real prospect of producing a much better situation.”

    No one expects mathematics to solve the problem to everyone's satisfaction. “It's ultimately a political problem,” Hirsch says. Kimball Brace, head of Election Data Services in Manassas, Virginia, and a member of the 2010 Census Advisory Committee, agrees. “Redistricting is contradictions out the wazoo,” Brace says. “One person's equality is another person's gerrymander.” Nonetheless, a growing group of practitioners believe mathematics can play a key role. Says Pildes, “Math can give you tools for creating processes that are likely to lead people to feel that the process is fair and that the outcome is therefore something to be respected.”


    Taking a Cue From Infinite Kinkiness

    1. Barry Cipra

    Mathematicians explored the mathematics of fractal billiards, in which a point-mass cue ball rattles around inside a shape whose boundary seemingly consists of nothing but corners, at the Joint Mathematics Meetings.


    “Clouds are not spheres, mountains are not cones, coastlines are not circles,” fractal pioneer Benoit Mandelbrot famously wrote. Pool tables, on the other hand, are polygons. But mathematicians can ask, “What if they weren't?”

    Robert Niemeyer, a graduate student at the University of California, Riverside, and his adviser, Michel Lapidus, are exploring the mathematics of fractal billiards, in which a point-mass cue ball rattles around inside a shape whose boundary seemingly consists of nothing but corners. Niemeyer displayed a few trick shots for future fractal hustlers at the Joint Meetings in a special session on experimental mathematics.

    “It's very cool,” says Victor Moll of Tulane University in New Orleans, Louisiana, who helped organize the session. “It's a new angle on a very classical, well-studied problem.”

    The research isn't purely recreational. Fractal billiards could account for details of what happens when sound bounces off a rough surface such as a seabed or the ceiling of a concert hall. Quantum physicists are also keen to understand billiard trajectories on all sorts of shapes for their connections with solutions to wave equations in a branch of physics called quantum chaos.

    Such applications are far in the future. For now, Niemeyer and Lapidus would be happy to find examples of periodic orbits—trajectories of finite length that wind up retracing themselves endlessly—on a familiar fractal known as the Koch snowflake. The Koch snowflake is formed from an initial equilateral triangle, onto which smaller equilateral triangles are repeatedly attached to the middle third of each side on the growing shape. In a paper published last year in The American Mathematical Monthly, Andrew Baxter of Rutgers University, New Brunswick, and Ronald Umble of Millersville University in Pennsylvania classified the periodic orbits on the equilateral triangle itself. Niemeyer hopes to parlay their results to classifications for successive approximations of the snowflake and ultimately to the limiting fractal.

    “There's a lot of interesting behavior when we look at these things analytically,” Niemeyer says. In particular, a fractal technique known as iterated function systems produces a promising set of trajectories that generalize the “Fagnano orbit” on the equilateral triangle, in which the cue ball simply bounces from midpoint to midpoint to midpoint. Some trajectories display a pleasing pattern of self-similarity, whereas others change radically on successive approximations of the snowflake (see figure).

    Corner pocket?

    Periodic orbits are easy to find inside this approximation of the fractal Koch snowflake but are much more challenging in the infinitely wiggly real thing.


    Fractal billiards fits well with the experimental approach to mathematics in which researchers use computers to systematically explore complicated phenomena and report the findings even when proofs remain elusive, Moll says. “I really like those pictures,” he adds. If the concept of bouncing off a boundary that's all kinks and sharp turns seems baffling, Niemeyer agrees: “It makes my head hurt, too.”


    The Joys of Longer Hangovers

    1. Barry Cipra

    Mathematicians at the Joint Mathematics Meetings showed a better way to get an overhang with a large set of bricks by translating the problem of stacking bricks into a problem about random walks.


    One of the joys of calculus is learning how to build a bridge to nowhere, by stacking bricks so that they extend as far out over the edge of a tabletop as you please. The key to the classic problem is that the “harmonic” series 1 + ½ + ⅓ + ¼ + … sums logarithmically to infinity (see figure). A “harmonic” stack of N bricks extends approximately ½ln N (ln being the natural logarithm). But a group of mathematicians has now left logarithms far behind, stacking bricks about as far as they can go.

    Balancing act.

    A “harmonic” stack of N bricks, each of unit length, overhangs a tabletop by half the harmonic sum (1 ½ + ⅓ + … + … + 1/N). Clever bricklayers can do better without ever getting their trowels dirty.


    In an invited address at the Joint Meetings, Peter Winkler of Dartmouth College described results that he and colleagues Mike Paterson of the University of Warwick, U.K., Uri Zwick of Tel Aviv University, Yuval Peres of Microsoft Research in Redmond, Washington, and Mikkel Thorup of AT&T Labs Research in Florham Park, New Jersey, have obtained on the best way to stack bricks. The work began with a “remarkable” and “completely new” result by Paterson and Zwick, Winkler says. In a paper just published in The American Mathematical Monthly, they showed that a better way to get an overhang with a large set of bricks is not to try extending things at each level but rather to build what looks like a parabolic brick wall with jagged edges. Even though it looks like a lot of wasted masonry, stacking the bricks that way boosts the overhang from a multiple of the logarithm of N to a multiple of the cube root of N—an unexpectedly huge increase.

    Building a wall to act as a counterweight for the overhang “hadn't occurred to people,” says Monthly editor Daniel Velleman of Amherst College in Massachusetts. “It's not obvious it's going to help. And the fact that it helps so much is surprising.”

    The question Paterson and Zwick left unresolved was how much better one might do. They were able to prove that no amount of cleverness can arrange N bricks to overhang by the square root of N. But that left a lot of room for possible improvement over the cube-root construction.

    Enter Winkler, Peres, and Thorup. In a follow-up paper slated for the Monthly, the five-man crew has nailed down 6N as an upper limit on overhang. They did so by translating the problem of stacking bricks into a problem about random walks. Adding a brick, it turns out, spreads force in essentially the same way that taking a random step spreads the walker's probability of being at a given location equally in each direction. How close to the multiple 6 it's possible to get, however, is unclear. Paterson and Zwick's parabolic construction achieves an overhang of 0.57N, and other constructions they have found suggest overhangs of about 1.02N.

    Other questions linger as well—the role of friction, for example. “Friction makes a real difference,” Winkler says. “And real bricks have a lot of that—not to mention mortar!”