News this Week

Science  16 Sep 2011:
Vol. 333, Issue 6049, pp. 1556
  1. Around the World

    1 - Washington, D.C.
    Senate Budget Bill Keeps DOE Science Level
    2 - Codolet, France
    Blast at Nuclear Waste Facility Kills One
    3 - Tilburg, the Netherlands
    Psychologist Sacked Over Faked Data
    4 - Harbin, China
    Lab Infection Rocks University

    Washington, D.C.

    Senate Budget Bill Keeps DOE Science Level

    On 7 September, the Senate Appropriations Committee approved $4.84 billion for the Department of Energy's (DOE) Office of Science in 2012. That's the same level as this year, and a slight bump over the $4.8 billion approved in July on a largely partisan vote by the House of Representatives covering the entire department. It's a far cry from the $5.416 billion the Obama Administration requested for 2012, but officials at the Office of Science's 10 national labs aren't complaining. In budgets, “flat is the new good,” quips Eric Isaacs, director of Argonne National Laboratory in Illinois.

    The Senate bill would slightly boost some programs, including basic energy sciences (BES) and biological and environmental research (BER). The BES budget would climb by $15 million, or 1%, to $1.693 billion, while BER funding would rise by $10 million, or 1.6%, to $628 million. The big loser in the Senate bill would be fusion energy sciences. Its budget would fall by $40 million, or 11%, from this year's level to $335 million.

    Codolet, France

    Blast at Nuclear Waste Facility Kills One

    An explosion at a plant that burns and melts nuclear waste killed one person and injured four others on 12 September but did not cause radioactivity to escape, French authorities say.

    The blast occurred in an oven at Centraco, a facility that turns low-level and very low-level radioactive waste into packages for storage. The plant, at a large nuclear site in the Gard department, is owned by Socodei, a daughter of energy giant EDF. The French Nuclear Safety Authority said the building was not damaged, the injured were not exposed to radiation, and there was no risk of environmental contamination.

    Tilburg, the Netherlands

    Psychologist Sacked Over Faked Data


    Diederik Stapel, a Dutch social psychologist and author of eye-catching studies about human behavior, lost his job at Tilburg University in the Netherlands on 7 September after the university concluded that some of the data in those studies were fabricated. Stapel is known as a prolific researcher whose studies appeared to offer new insights into the workings of the human mind; for instance, a Science paper published in April showed that people are more likely to stereotype or discriminate in messy environments.

    University Rector Philip Eijlander said in a TV interview 7 September that he was first contacted on 27 August by “junior researchers” in Stapel's lab who alleged that his conduct was fraudulent. Stapel immediately admitted that there were “strange things” in his papers, Eijlander says, and on 6 September, that some data were faked. The university has asked Willem Levelt, a psycho linguist and former president of the Royal Netherlands Academy of Arts and Sciences, to lead a panel investigating the extent of the alleged fraud as well as the research culture at the university that apparently enabled it. Eijlander says that all “tainted papers” will be retracted.

    Harbin, China

    Lab Infection Rocks University

    China's Northeast Agricultural University sacked two administrators and announced it would compensate students infected in a lab outbreak last December, acknowledging that the school did not take sufficient safety precautions.

    Twenty-seven veterinary students and one instructor contracted brucellosis, a bacterial infection that can cause fever, weakness, and muscular pain, after dissecting goats that had not been properly quarantined before reaching the lab. In an interview with the newspaper Southern Weekend, an infected student alleged that an instructor discouraged the group from wearing gloves when handling the animals.

    China has spent billions of dollars creating a handful of world-class laboratories and universities. But conditions at lesser institutions have suffered, says Gerry Postiglione, an education scholar at the University of Hong Kong. Many have “cut corners on cost and quality,” he wrote in an e-mail. Along with dismissing the dean and the Communist Party secretary at the College of Veterinary Medicine, the university offered 61,000 yuan ($9545) per student in tuition waivers, medical fees, and compensation. The incident has focused national attention on the issue of lab safety, prompting calls for reform.

  2. Newsmakers

    Planetary Scientist Admits Attempted Espionage

    Stewart Nozette, a one-time principal investigator for NASA's Lunar Reconnaissance Orbiter, admitted 7 September that he had tried to sell classified materials to an FBI employee posing as an Israeli intelligence agent.

    Nozette was already in trouble with the law when he fell for the FBI sting. In January 2009, Nozette pled guilty to overbilling the government by $265,000 on government contracts awarded to a technology company he ran. While he awaited sentencing, Nozette met with the supposed Israeli agent in October 2009, having received $11,000 in return for passing on classified materials about U.S. satellite defense systems. In a videotaped conversation in the hotel, he negotiated for a new identity, a passport, and travel to another country. Apparently, Nozette, 54, was ready to give up his life as a leading planetary scientist in the search for ice on the moon. “I've crossed the Rubicon,” he said in the hotel room. “I've made a career choice.”


    Lasker Awards for Protein Folding, Malaria Drug

    Two scientists who discovered a cell “machine” involved in protein folding and a researcher who found a life-saving malaria drug are this year's recipients of the Albert and Mary Lasker Foundation's Lasker Awards. The winners receive a $250,000 honorarium.

    Biochemist Franz-Ulrich Hartl, 54, of the Max Planck Institute of Biochemistry in Martinsreid, Germany, and biologist Arthur Horwich, 60, of Yale University share this year's prize for basic medical research. In the late 1980s, they discovered that within the cell, a cagelike molecular apparatus dubbed a chaperonin helps a linear string of amino acids fold into its correct three-dimensional shape.

    The 2011 Lasker Award for clinical research goes to Tu Youyou, 81, of the China Academy of Chinese Medical Sciences, Beijing. During China's Cultural Revolution, Tu took part in a project to find a treatment for chloroquine-resistant malaria parasites. Her team tested hundreds of traditional Chinese herbal extracts in malaria-infected mice, finally settling on one made from sweet wormwood. From this extract they purified artemisinin, a drug that has saved millions of lives.

    A third prize for public service went to the National Institutes of Health's Clinical Center.

    Cancer Patients Sue Duke University

    A dozen plaintiffs have filed a lawsuit against Duke University and a handful of its administrators, researchers, and physicians, alleging that the defendants engaged in fraudulent and negligent behavior when they enrolled cancer patients in clinical trials based on faulty data. The lawsuit, filed 7 September in a North Carolina court, comes 14 months after Duke oncologist Anil Potti admitted he had embellished his resume; he later resigned (Science, 6 August 2010, p. 614). A raft of papers Potti co-authored with Duke cancer geneticist Joseph Nevins were retracted. The plaintiffs—cancer patients and their families—say Duke officials knew the work was “highly suspect” but launched the clinical trials anyway. The trials assigned treatment based on now-discredited gene expression patterns that Potti and Nevins said they had identified.

    Duke's response “to the accusation of invalid and fraudulent science was deceptive, misleading, and fraudulent conduct,” the lawsuit reads. The plaintiffs are seeking damages and a trial by jury.

    Cosmic Microwaves and Quantitative Genetics Win Balzan Prizes

    An astrophysicist and an evolutionary biologist are among the winners of this year's Balzan Prizes, awarded by the Milan-based International Balzan Foundation.

    Joseph Silk, a professor at Johns Hopkins University in Baltimore, Maryland, will receive the prize for his contributions toward understanding the early evolution of the universe. Silk has studied how the initial fluctuations in the cosmic microwave background helped to shape the first galaxies. Russell Lande, a professor at the University of California, San Diego, receives the honor for helping to elucidate the relationship between genes and quantitative traits such as body size in animal populations. Lande's work is now a cornerstone of theoretical population biology. Each Balzan prize awardee receives $950,000. Winners are obligated to allocate half of the prize money to research projects, preferably involving young researchers.

  3. Random Sample

    The Sky Is Falling! (But No Big Deal)


    NASA announced on 9 September that the Upper Atmosphere Research Satellite (UARS), sent into orbit in 1991, is coming back down by the end of this month. Where the remains of UARS will land is unknown, but NASA calculates that the chances of UARS debris hitting a particular human are about 1 in 22 trillion, based on how UARS will likely break up as it hits the atmosphere and the shape and composition of those pieces. Some pieces will burn up; for the 26 surviving pieces, it's a question of how far a piece will travel and what there is to hit (which leaves out the 70% of the planet that's water and everything poleward of 57° latitude, over which UARS does not orbit).

    Even 2 hours before reentry, there will still be so much error in the prediction of the 800-kilometer-long debris-strewn field that it could be anywhere along a 10,000-kilometer-long track. Getting hit is unlikely, but if you are anywhere near that reentry track, it should be quite a sight: a $750 million fireworks display.

    By the Numbers

    1.46 — Kilograms of carbon dioxide generated by a typical Google user per year, according to the company.

    £1.6 billion — Amount that science funding in the United Kingdom will decline by 2014–15, according to an analysis released 14 September by the Campaign for Science & Engineering.

    Search and Rescue, Without the Dogs


    The nose of a German shepherd in a convenient, carry-on-sized backpack? Coming soon, says Paul Thomas, a chemist at Loughborough University in the United Kingdom. Thomas was inspired to develop toteable life sensors for search and rescue teams after speaking with a member of the U.S. National Guard who led a team of searchers locating people trapped in the wreckage of New Orleans after Hurricane Katrina ravaged the city in 2005 (pictured right). After 8 hours of hunting, Thomas says, the guard's team was exhausted.

    “Dogs get tired. People get tired,” Thomas says. “You need something that doesn't get tired.” So Thomas, with colleagues from the international project Second Generation Locator for Urban Search and Rescue Operations, set out to create devices that could locate people trapped beneath rubble using tell-tale signs of life such as carbon dioxide and ammonia emitted from the body. Study subjects breathed and sweated for hours—carefully monitored—within a sarcophagus-shaped box in a laboratory (at left). Meanwhile, these gases passed out of the box and into a long cylinder that filtered them through simulated wreckage resembling a parfait of smushed layers of concrete, insulation, and other construction materials.

    Even through the faux rubble, an array of sensors picked up the subjects' chemical distress calls, Thomas and colleagues report this week in the Journal of Breath Research. Many of those sensors are portable, too—suggesting the team can, indeed, squeeze the business end of a rescue dog into a backpack.

  4. Scientific Community

    Particle Physicists' New Extreme Teams

    1. Adrian Cho

    Life at the world's biggest atom smasher is an odd combination of selfless cooperation and intense competition.

    In your face!

    A life-sized photo of CMS hangs across from the ATLAS team's offices.


    MEYRIN, SWITZERLAND—Eighteen years ago, as an undergraduate student at Eindhoven University of Technology in the Netherlands, Martijn Mulders worked on an experiment seemingly ideal for a physicist in training. Using lasers, he would study fluctuations in a glowing plasma, work directly relevant to the manufacturing of microchips. The small-scale “tabletop” experiment gave Mulders control over every aspect of the work.

    Yet he found the experience wanting. “With a tabletop experiment, it's just you and the tabletop,” Mulders says. “You are kind of isolated.” So, as a graduate student at the University of Amsterdam, he switched fields, moved here to the European particle physics laboratory, CERN, just west of Geneva, and did his thesis research as one of 550 members working with a particle detector called DELPHI. “What I really like about life in a big collaboration is that there's always plenty of challenging things to do,” Mulders says. “What you do gets appreciated.”

    Mulders, 39, now works in perhaps the biggest scientific collaboration ever assembled. Three years ago, CERN turned on the world's highest-energy atom smasher, the Large Hadron Collider (LHC). Aiming to create new particles and maybe even open new dimensions, the circular accelerator blasts protons together within four huge particle detectors spaced around its 27-kilometer circumference. The two largest detectors, known as ATLAS and CMS, vie for those discoveries, while one called LHCb studies certain familiar particles in great precision and another called ALICE studies a form of nuclear matter produced when the LHC smashes lead ions. Some 3000 researchers work on ATLAS, and 3600 work on CMS, including Mulders. “The whole world of particle physics is here,” he says. “So I don't look at it as a big collaboration but as a small world. … It's perfect being here.”

    It's something of a brave new world for particle physicists. Since their field was born in the 1930s, they have worked on ever-bigger machines in ever-bigger teams. For decades, collaborations of hundreds of researchers have been the norm. But by pushing into the thousands, the two large LHC collaborations confront physicists with new issues and pressures. That's especially true because the LHC will soon be the world's sole great atom smasher, leaving ATLAS and CMS with only each other for competition.

    With fewer rival teams and many more teammates, researchers working on ATLAS and CMS face as much competition from within the collaboration as from without. Particle physics relies on an extreme division of labor, but scientists now face the reality that, even as they work on their specialized tasks, others within the same team are doing the same thing. In the past, one collaboration worried mostly about getting scooped by another; now members of a collaboration seem to worry as much about getting scooped by their own teammates.

    “In the old experiments, when we had 300 or 400 people, really it was an easier job,” says George Mikenberg, an ATLAS member from the Weizmann Institute of Science in Rehovot, Israel, who has worked at CERN since 1982. “You can remember 300 faces.” Still, he and other physicists say they're happy in the collaborations of thousands. “This works,” Mikenberg says, “so what the hell?”

    Among the worker bees

    CERN's building 40, which houses the ATLAS and CMS collaborations, feels like a gigantic beehive. Within the eight-story cylindrical structure, balconies of cubicles ring a vast atrium. On the ground floor, a café serves physicists who gather in twos and threes, their conversations melding into a multilingual thrum. Puckishly, researchers with CMS have plastered a life-sized photo of the detector on their side of the atrium. The ATLAS team can't respond in kind because a full-scale image of 25-meter-tall ATLAS won't fit.

    A particle physics collaboration works a bit like a Woodstock-era commune. Members cooperate in running and maintaining their detector, collecting the raw data, and converting them into a readily analyzed format. Later, like so many commune members gathering for dinner, they go through the data, breaking into smaller groups to search for particular particles or phenomena—just as some commune members might opt for the nut loaf while others prefer the tempeh. Once the collaboration has approved a result for publication, essentially all members put their names on the author list.

    “I don't look at it as a big collaboration, but as a small world. … It's perfect being here.”



    Also like a commune, most particle physics collaborations are run as vaguely defined democracies. Both ATLAS and CMS have a few elected officers, a hierarchy of boards and committees, a slate of working groups to help manage the work—and little means to force anybody to do anything. “The political system is quite close to anarchy—not in the pejorative sense, but in the sense that there is very little formal authority,” says David Coté, a postdoc at CERN who works on ATLAS.

    In more detail, each detector is a high-tech canister surrounding a point at which the LHC's countercirculating beams collide. Those collisions can blast into existence massive new subatomic particles that quickly decay into telltale combinations of familiar ones. The detector's myriad subsystems aim to characterize that debris. So for “service work,” a physicist might help calibrate the subsystem that detects particles called muons or tune the “trigger” software that identifies promising collisions and makes the detector record its data.

    A similar division of labor holds sway in the sexier arena of data analysis. For example, ATLAS and CMS physicists are in hot pursuit of the long-sought Higgs boson, the key to their theory of how all the other particles get their mass. To find it, they look for combinations of particles that theory says the Higgs could decay into: a pair of photons; two massive particles called W bosons; or a particle called a bottom quark and its anti-matter partner, among many other possibilities. So a Higgs hunter might join a group that is focusing on just one decay mode and work on some detail of that analysis.

    Coaxing researchers to do their share of service work is a perpetual challenge, physicists in both collaborations say. “I think it's a constant struggle in all experiments, not only ATLAS, that you have this tension between the low-level work and analysis,” says Sara Strandberg of Stockholm University, who heads ATLAS's “combined performance group” to check the quality of the data coming out of the detector.

    Even with thousands of teammates, life in the collaborations is homier than it appears from the outside, physicists say. Coté is searching for new particles predicted by a concept called supersymmetry, or SUSY. So are the 500 other physicists in ATLAS's SUSY working group, he says. But only a fraction of them work in the subgroup that covers the SUSY signature Coté is working on with a couple dozen colleagues. And on a daily basis he works with a few colleagues at CERN. “We are a little core group of four people who are interacting every day,” Coté says.

    A crowded table

    The huge LHC collaborations continue a decades-long trend toward bigger teams in particle physics. From 1989 to 2000, CERN ran the Large Electron-Positron collider, which fed four detectors, each with a team of hundreds of physicists. At Fermi National Accelerator Laboratory (Fermilab) in Batavia, Illinois, the Tevatron collider smashes protons and antiprotons at one-third the LHC's energy and feeds detectors called CDF and D0 that support teams of 500, down from 700 a few years ago. (The 25-year-old Tevatron will shut down later this month.)

    The sheer scale of the ATLAS and CMS collaborations, however, brings new factors into play. “The increase in size from CDF and D0 to the LHC collaborations was a factor of 6,” says Shahram Rahatlou of Sapienza University of Rome, who heads a CMS working group searching for exotic new particles. “The number of physics topics you can study has not increased by a factor of 6. It's maybe a factor of 2.” So what do you do with six times as many people?

    The answer: Compete with your neighbor. Some of CMS's 10 physicsanalysis working groups are now as big as the current CDF and D0 collaborations, Rahatlou says. And within those working groups, every interesting analysis will be pursued independently by at least two teams, sometimes several more. The same holds true in the ATLAS working groups.

    Such competition produces invaluable crosschecks, physicists say. “It's absolutely essential for topics as important as the Higgs and SUSY that you have multiple teams working in parallel,” says John Ellis, a theorist at CERN, who is not a member of either collaboration.

    Still, researchers can have too much of a good thing. When the LHC started taking data in March 2010, scientists set out to remeasure the properties of known particles, such as the massive top quark, which had been discovered at the Tevatron in 1995. In both collaborations, four or five teams measured basic parameters such as the top quark's mass. And in each collaboration, leaders of the top quark working group had to choose one analysis for publication. “You couldn't even use one as a crosscheck of another because they were practically identical,” says Christophe Delaere, a CMS member from the Catholic University of Louvain in Belgium.

    Physicists say that experience has spurred them to be more creative in devising analyses to avoid overlap. All agree that if people are willing to do what's needed, there's plenty of work to go around. “If you ask any physics analysis group in ATLAS, they will tell you they need more people, even with 3000 of us,” Strandberg says.

    Still, on hot topics, toe-to-toe competition seems inevitable. In July, both ATLAS and CMS reported at a conference possible hints of the Higgs boson, especially as it decays into two W bosons (Science, 29 July, p. 507). Six different groups within ATLAS now want to improve the Higgs-to-WW analysis, says Bill Murray of Rutherford Appleton Laboratory near Didcot, United Kingdom, senior convener of the ATLAS Higgs working group. “We can't publish all six different analyses,” Murray says. “We have to pick the best one.” The others will be written up in internal documents, he says.

    Look out below!

    “If you ask any physics analysis group in ATLAS, they will tell you they need more people, even with 3000 of us.”



    Even amid extreme competition, researchers say, most collaboration members play by the rules and work with working groups. But with Nobelcaliber quarry like the Higgs boson and SUSY particles finally within range, physicists worry that some may opt for rougher tactics. “There is the good-citizen approach, and then there is the approach ‘I am better than you, and I'm going to kill you,’ ” says Maurizio Pierini, a postdoc at CERN and a member of CMS.

    Physicists particularly worry that a few rogue researchers might follow along on a hot topic and then swoop in at the last moment with their own version of the work, presenting the collaboration or working group with a fait accompli to claim discovery. “There was a rule even before we started taking data that this kind of parachuting is not allowed,” says Maria Spiropulu of the California Institute of Technology in Pasadena, who works on CMS. But all collaboration members have access to the data, so there's little to prevent someone from trying.

    In fact, some physicists say such a ploy has already been attempted. In April, blogs and other news sources buzzed with reports that an ATLAS team led by Sau Lan Wu of the University of Wisconsin, Madison, had spotted the Higgs boson decaying into two gamma rays. Outraged researchers sensed an attempt to scoop the collaboration and scrambled to find the source of the leak. According to blog accounts, somebody left an internal ATLAS memo describing the analysis on a printer, and somehow it found its way to the public. Murray, the senior convener of the ATLAS Higgs working group, says he became aware of the group's analysis only when he read the memo. “They didn't even let me know they were working on it,” he says.

    Wu denies that she was trying to get the drop on the collaboration. Her team had presented a preliminary search with much less data to a subgroup of the ATLAS Higgs working group in November 2010, she says. When she saw the signal in April, the subgroup was not scheduled to meet for another week or so, so her team wrote the internal memo, she says. The memo was an attempt to follow the rules, Wu says: “If there had not been a leak, this would not have been an issue.” Wu insists her team did not leak the signal—which proved spurious—and says somebody else may have done so to discredit her.

    In the wake of the incident, both collaborations moved to force people to work within the working groups, researchers say. And most physicists say rogue researchers have little hope of besting an analysis that has received ongoing attention from the working group. That's especially true because to make it to submission for publication, an analysis must first win the approval of the working group. The details differ between ATLAS and CMS, but after that step an analysis must be presented to the full collaboration both in a talk and in a preprint. Researchers must respond to all comments before leaders release the paper. (After such vetting, journal peer review is often quick.)

    Nevertheless, it might not be so easy to turn a blind eye to a rogue analysis, says Christoph Paus of the Massachusetts Institute of Technology (MIT) in Cambridge, who is co-convener of the CMS Higgs working group. After all, the main competition is still the other collaboration. “If somebody comes and has a very good analysis that's almost complete, you can't just ignore it,” Paus says.

    Murray says that the April incident has raised the only allegations of such tactics he's aware of. Still, physicists can estimate when certain analyses should pay off. For example, within a year, researchers should have enough data to either spot the Higgs boson or rule it out. As fruition nears, somebody else could well be tempted to parachute to glory.

    Two cultures

    In the L-shaped control room for the CMS detector, banks of computer monitors cluster in three islands. Overhead, more monitors show graphs that keep tabs on the detector and its 100 million data channels, the trigger that makes it record events, the hardware that does the recording, and many other facets of the machine. Occasionally, a snatch of a song—a bit of funk or a measure of U2's “In a Little While”—pierces the chatter of a dozen researchers taking “shift” to alert them to one condition or another.

    The room has a thrown-together look. The shift leader—Delaere of the Catholic University of Louvain—explains that CMS researchers had planned to put computers in this room and the control room in a larger one above it, complete with a broad window through which observers could look. But then physicists decided they needed more computers, so the machines wound up in the bigger, brighter room and the people in the smaller, danker one.

    The story hints at the cultural differences between CMS and ATLAS. “ATLAS somehow I think of as being very Swiss,” says CERN theorist Ellis. “My impression is that it's a very democratic collaboration that has a very well-framed constitution and rules—sort of what you'd expect from the Swiss.” CMS, Ellis says, is “more of a seat-of-the-pants operation.”

    Others say that ATLAS management strives more for consensus whereas CMS management is more “top down.” “CMS people complain that decisions are taken at a high level and that it affects the work you're doing at a lower level,” says CMS member Pierini. “ATLAS people complain that there aren't any decisions at a high level and that this leads to confusion on the lower level.”

    That difference has practical consequences, physicists say. When data-taking began at the LHC in March 2010, Pierini and colleagues missed a deadline to install a trigger setting they needed for their SUSY search. Unable to persuade CMS-run management to extend the deadline, they missed out on the first month of data-taking. “A little flexibility would have let us put the trigger in a day later,” Pierini says.

    ATLAS's populist approach also has its drawbacks. For example, ATLAS has two software systems from two different groups for tracking muons. That's because the collaboration had no way of telling one group to yield to the other or of making the groups work together. “ATLAS is definitely willing to duplicate effort,” says Joshua Cogan, a graduate student from Stanford University in Palo Alto, California. “And ATLAS is definitely willing to let you pursue something even if they know [the collaboration] will veto it” in the end.

    The view from the bottom

    Within the gleaming ATLAS control room—which, with its slick, curving consoles and gigantic computer displays, looks like the set of a network newscast—an alarm pings. It alerts the ATLAS shift leader that, with data-taking paused, a physicist in the United States wants to test one of the detector's many subsystems. With a click of a mouse, the shift leader transfers control of the “hadron calorimeter,” demonstrating how ATLAS and CMS can be controlled from around the globe.

    “There is the good-citizen approach, and then there is the approach ‘I am better than you and I'm going to kill you.’”



    Yet for all the power of decentralized control, the collaborations seem to be driven by humbler bits of technology: coffeemakers. They are everywhere at CERN, many of them the fancy jobs that grind their own beans and make an espresso good enough to satisfy Italian researchers. Each one is emblematic of the importance of being here.

    Meeting for coffee “is superimportant,” says Thilo Pauly, a postdoc at CERN who coordinates the daily operations of ATLAS. “In order to make decisions, you have to go by the normal procedure and go to the right meetings,” he says. “But of course, it helps if you start to make alliances, … and the preparation is typically done in coffee meetings.” Stephanie Majewski, a postdoc at Brookhaven National Laboratory in Upton, New York, says she didn't drink coffee until she came to CERN. “You meet with eight different people and have eight cups of coffee, and by the end of the day you're shaking,” she says. Like all the other young physicists interviewed by Science, she says that a stint at CERN is essential for getting ahead.

    Those at the bottoms of the ATLAS and CMS heaps, the postdocs and graduate students, say the competitive pressures can be intense. “Very quickly it can become very demotivating to think that there are 150 other people working on the same thing you are,” says Maiken Pedersen, a graduate student from the University of Oslo who is working on a SUSY search on ATLAS. “It's difficult to be fast enough” to compete with others. And there are many others: ATLAS and CMS employ about 1000 graduate students and hundreds of postdocs each.

    So how does a young physicist stand out in the crowd? “Shameless self-promotion,” ATLAS's Majewski says with a laugh. “When writing up [a job] application, you really just have to brag about yourself, which some physicists are pretty good at.” Majewski worries that in the huge collaborations, people with better social skills may rise faster and further than those with the better scientific skills.

    Others say that collaboration members know who has done what and that talent wins out. “In the end, the right people are getting the jobs,” says MIT's Paus, the co-convener of the CMS Higgs working group. “Not everybody can stay in the field.”

    Be that as it may, young physicists say they feel fortunate to participate in such a grand adventure. “We may be able to make discoveries that will alter the future of physics,” says Benjamin Hooberman, a postdoc from Fermilab who works on CMS. “And I want to be a part of that.” That spirit keeps 3000 physicists working and playing (fairly) well with one another in an environment in which self-interest and team interest seem to collide about as often as the protons whizzing around at light speed.

  5. Scientific Community

    U.S. Physicists, a Long Way From Home

    1. Adrian Cho

    The United States has gone to great lengths to keep its scientists at the world's biggest atom smasher, the Large Hadron Collider, integrated. Nevertheless, many make personal sacrifices to be there.

    MEYRIN, SWITZERLAND—It's not hard to find an American here at the European particle physics laboratory, CERN. Among physicists working on the particle detectors fed by the world's biggest atom smasher, the Large Hadron Collider (LHC), researchers from the United States outnumber those from any other nation. Of the 3600 researchers working on the massive CMS detector, 900 hail from the United States, as do 700 of the 3000 researchers working on the ATLAS detector.

    Still, working at CERN isn't like working at home, says Aaron Dominguez, a CMS member from the University of Nebraska, Lincoln. The United States is not one of CERN's 20 member nations, so U.S. researchers have little say in how the lab is run. “You learn how the lab works, and you deal with it,” Dominguez says. “It's the role of a guest.”

    The United States has gone to great lengths to keep its scientists integrated in the far-away experiments—for example, by establishing a remote center for CMS at Fermi National Accelerator Laboratory in Batavia, Illinois. Nevertheless, many make personal sacrifices to be here. Vivek Sharma of the University of California, San Diego, is co-leader of the working group within the CMS team that's searching for the Higgs boson. He spends 8 weeks at CERN for every week at home with his wife and their 7-year-old daughter.

    “It's more of a sacrifice for them,” Sharma says. On weekends, he says, he and his family rely on Internet video links to “be” together: “When they wake up, we just put on the cameras. They go about their things and I go about mine, and we have conversations.”

  6. Personalized Medicine

    Pushing the Envelope in Neuroblastoma Therapy

    1. Jennifer Couzin-Frankel

    A study in children with few options is matching drugs to their tumors and creating mouse models for each patient. Whether the science is up to it is a matter for debate

    Seeking a match.

    Identifying drugs that disable features of specific tumors is a dream of cancer researchers.


    Five years ago, doctors told Patrick Lacey that his son Will was going to die. Will had neuroblastoma, a cancer that forms in nerve tissue in young children; half are diagnosed when the disease is advanced, and most don't survive it. “I'm looking at my only child, he's 2 years old, and they're saying there's nothing we can do,” says Lacey, a bear of a man with reddish-brown hair and a Boston accent. “That's an unacceptable answer.”

    Lacey started networking. He reached out to families whose children, like Will, had relapsed. He scoured the scientific literature in search of therapies. And one day, at a gathering in New York City in 2007, he met Giselle Sholler, a pediatric oncologist then at the University of Vermont.

    Lacey was struck by how Sholler, a young doctor who had only recently completed her training, reacted when he told her about Will's dismal prognosis. “Unlike everyone else, she said, ‘Why is he incurable?’” Lacey recalls. “These kids have been incurable historically,” Sholler told him. “But that doesn't mean they'll be incurable forever.”

    Like every doctor treating neuroblastoma, Sholler knew that standard therapies were unlikely to cure advanced disease. In part because there is substantial genetic variation from one neuroblastoma tumor to another, the cancer frequently evades treatment. Once relapse hits, as it does for two-thirds of high-risk patients, survival becomes a dream out of reach. About 1% are cured.

    Sholler's vision is the same as that of most oncologists today: personalize care so that the drugs given to patients target specific features of their tumors. This is already beginning in neuroblastoma, as pediatric oncologists focus on a relative handful of genetic mutations and other vulnerabilities in tumor cells that respond to experimental or recently approved drugs. But Sholler wanted more, and she wanted it quickly. The science of matching drugs to tumors, she knew, was imperfect. But her patients couldn't wait.

    Lacey, taken by Sholler's eagerness to tackle neuroblastoma from any angle, began raising money to support her. Foundations run by him and other neuroblastoma families have donated hundreds of thousands of dollars to her lab, making possible Sholler's most unorthodox venture yet. She and colleagues are surveying tumor gene expression patterns in children with dwindling options and using algorithms to match tumor features with available drugs—including drugs not designed to treat cancer. Sholler's group is also creating mouse models of every child's cancer and testing therapies on the animals, to help determine whether the matching system is on target.

    Will isn't in this trial. The first to enroll was another boy, a 12-year-old who's battled neuroblastoma for more than half his life. He was treated last month at the National Cancer Institute (NCI) in Bethesda, Maryland, with a combination of three drugs, two of which wouldn't normally be on anyone's radar screen. The study reflects what many say they want to do in cancer treatment but few have dared to try. There's a reason for that, some oncologists caution: Science and mathematics just aren't solid enough to make reliable tumor-drug matches consistently.

    That hasn't stopped parents—many of whom are almost as scientifically savvy as the doctors caring for their children—from tirelessly raising money and volunteering to sign up their children. “I am not expecting every single kid who enrolls in this [trial] to be cured, but what I am expecting is the data that's collected will give a much better sense of what's going on with these kids' tumors,” Lacey says. “Someone has to do it first.”

    Drugs by algorithm

    When Lacey first met Sholler in New York 4 years ago, her research was just getting off the ground. She had started an unlikely clinical trial of a drug called nifurtimox: It is approved to treat Chagas disease, a parasitic disease of the tropics. A patient of Sholler's with neuroblastoma had contracted Chagas from a blood transfusion and was prescribed nifurtimox, the standard treatment. To everyone's great surprise, the little girl's tumors shrank. Subsequent analysis showed that nifurtimox killed neuroblastoma cells in the lab.

    Sholler launched a small trial of nifurtimox, which has since advanced to a larger phase II study. She subsequently added several other early-stage trials to her list, including one of a drug called DFMO, which is used to treat African sleeping sickness (also a parasitic disease) and excessive hair growth, but is also being tested experimentally in some cancers.

    But Sholler's eye was really on drugtumor matching. There are numerous road-blocks. One is time: Children with relapsed neuroblastoma normally need treatment every few weeks, so researchers don't have the luxury of spending months hunting for the best-suited drugs. Another is accessing the tumor for study. Neuroblastoma is often hidden deep in the body; gathering tissue can be invasive, even risky, especially for children whose bodies are already battered by disease and treatment. Most importantly, there's no consensus on how to identify the best drugs.

    Several groups are honing selection methods. One, led by Atul Butte, a pediatrician and bioinformaticist at Stanford University in Palo Alto, California, published two papers last month in Science Translational Medicine ( and detailing a strategy to identify existing drugs that might help treat diseases for which they weren't designed. Butte and his colleagues combed data repositories for gene expression changes associated with various diseases and for drugs that counter those genome-wide shifts. An ulcer drug, they theorized, might help lung cancer; a seizure drug, inflammatory bowel disease.

    Funding source.

    Driven to save his son Will, Patrick Lacey's fundraising, and that of other neuroblastoma families, has made Giselle Sholler's research possible.


    To get their matches, they relied on an algorithm to find cases for which the gene expression change in a disease, and the genome-wide shifts wrought by a drug, were strong enough statistically to counteract each other. “You never get perfect matches here,” Butte says, but “my whole attitude is never to wait for perfection” before trying something new in patients. Still, he says his findings are tentative; he doesn't want people with lung cancer running out to the drugstore and swallowing ulcer pills before more is known about their potential effects.

    Some links might be tentative but not all, Sholler believes. She and others wanted to forge ahead. Early this year, she was lured from Vermont to the Van Andel Institute, a research center in Grand Rapids, Michigan, where researchers a few years earlier had completed a pilot trial matching a hodgepodge of drugs to 50 different cancer patients. Most were adults. The goal was relatively simple, “just to ask the question, can we do this in real time?” says Craig Webb, the Van Andel molecular oncologist and bioinformaticist who helped lead the study and developed the technology they used. For example, in a 14-year-old girl with leukemia, tumor cells had very high levels of a protein called FLT3. The model flagged a kidney cancer drug called sorafenib, which was then still experimental, as targeting the protein. So doctors offered it to the teenager. Although her cancer improved, she later died from complications of a bone marrow transplant.

    Still, Sholler was impressed when Webb shared the trial's response rate. The effect was variable, but doctors treating these patients believed that about one-third showed some improvement in their cancer. Sholler says that's three times greater than what many early-stage trials record in people as sick as these. Webb hasn't published the results in full: The algorithms kept evolving with each patient as he and his colleagues worked to refine them. “You can imagine trying to write that up in a methods section,” Webb says.

    Since then, Webb has kept tinkering. The strategy has stayed the same: Look for a gene that's abnormally expressed and a drug that can target its protein product or another in the same pathway. But his group is trying to focus more on the science, trying to understand what might be causing a gene to behave wildly in the first place, and treat accordingly. “We do anchor this in the literature,” he says. “It's not just a blind leap of faith.”

    One source of new data is the Connectivity Map being developed at the Broad Institute in Cambridge, Massachusetts. Researchers there have been treating cancer cells in the lab with hundreds of potential therapies, to analyze their effects on the cells' gene expression patterns. So far, they've experimented with about 3000 compounds and generated 150,000 different gene expression profiles. The Broad Institute posts results on a public database, enabling Webb, Butte, and others to incorporate the information into their models.

    With genetic technology moving fast, some wonder whether it would make sense to go straight to DNA sequencing rather than rely on gene expression to analyze tumors. Hundreds of papers in the last several years have proposed expression “signatures” that correlate with prognosis or response to particular treatments; few have been replicated and even fewer have made it to the clinic.

    Neuroblastoma specialists are watching the Van Andel trial with interest but also with skepticism. One who's waiting for stronger data is John Maris, chief of the oncology division at Children's Hospital of Philadelphia. “It's not as simple as, you take a tumor, you throw it on an array, and you have a printout of what drugs will work,” he says. Maris has been working on a targeted therapy that's designed to help patients with ALK mutations in their tumors; it's currently in clinical trials. “What Dr. Sholler is trying to do is exactly what I want to do, what all of us want to do,” says Maris, who considers Sholler and her colleagues friends. But he worries that we haven't yet discovered the Achilles' heels in cancer cells that are “real and targetable.”

    Most of all, Maris worries that Sholler's trial will give false hope to families desperate to save their children—hope that this treatment is truly personalized, when he's not sure it really is.

    Extreme medicine

    Families may be frantic to try anything. But doctors involved in the Van Andel study say that, as clinical trial veterans, parents know full well that a response to treatment is far from guaranteed. “I am very upfront with the patients,” says Melinda Merchant, the pediatric oncologist at NCI who treated the 12-year-old boy who took the trial's first slot. While doing more than gene expression might be ideal, Merchant says, sequencing the “exome”—the DNA that codes for proteins—takes at best 4 weeks, time that neuroblastoma patients don't have. The next stage of the study is slated to include some sequencing, however.

    Wily foe.

    Neuroblastoma tumors are genetically heterogeneous, which helps make them elusive targets. The disease strikes young children and those with high-risk disease often don't survive.


    In addition, “we are bringing the patient's tumor to the lab” and injecting the cells into mice, says Sholler, who treats patients at the Helen DeVos Children's Hospital across the street from the Van Andel. She hopes there will be a mouse model for every patient enrolled in the trial. Researchers will use the mice to test the drugs recommended by the algorithms—along with those it recommended against—to determine whether they're on the right track.

    In part to get others on board, Sholler was prepared to jump through numerous hoops before opening up the trial. She and her colleagues submitted their protocol to the U.S. Food and Drug Administration; they also submitted the algorithms for FDA approval as a “device.” FDA okayed it. The trial opened at the end of June with five centers that can enroll patients.


    Diagnosed with neuroblastoma as an infant, Will Lacey celebrated another birthday last month.


    Once a tumor is analyzed and drugs proposed, a “tumor board” of about two dozen members meets to discuss the options. They consider the scientific rationale for each drug and the safety of combining them; up to four can be given to a patient. One hundred thirty-one different drugs that offer guidelines for pediatric dosing are eligible; about half are cancer drugs. From start to finish, the process takes 2 weeks. “The primary goal is safety for the patients,” Sholler says. “It really is a very early trial.” It has room for 14 patients, and Sholler expects to have results within a year.

    Funding is a challenge. Sholler applied to the Department of Defense, which supports some cutting-edge neuroblastoma research. But “it was just too new, too uncertain,” and the proposal was rejected, she says.

    The tiny handful of other researchers doing similar work can relate. At the Translational Genomics Research Institute (TGen) in Phoenix, Arizona, physician-in-chief Daniel Von Hoff is running a trial in refractory breast cancer similar to Sholler's, but he's focused solely on “deep” sequencing: hunting for mutations and other aberrations across the genome, as opposed to gene expression patterns. He published a pilot version of his trial last November in the Journal of Clinical Oncology. Eighty-four patients with various forms of advanced cancer participated. About a quarter fared better on the new regimen, designed to target their tumors, than they had on the previous one.

    Von Hoff 's money has come from industry and private foundations. “[I] haven't applied to NIH [National Institutes of Health],” he says. “I'm kind of afraid to apply there, because they'll say it's probably not ready yet.”

    Sholler has relied heavily on the fundraising efforts of Lacey and other neuroblastoma families. In all they've raised more than $1,750,000 over 3 years, $400,000 of which will go to her latest trial. About $1,000,000 of that has come from two foundations Lacey runs. “People aren't just going to start funding federal grants … because some parents say, ‘This looks exciting,’” Lacey says.

    He holds bake sales, raffles, live auctions, and, one year, an elaborate Thanksgiving party to raise money. Lacey formed his own foundation, Friends of Will, after an organization in which he was involved evaporated when the other families running it lost their children to the disease. Fundraising has “taken an incredible toll on me personally,” Lacey says. But at the same time, he sees the end result: A fundraiser one year paid for preclinical work on the drug DFMO; the next year, Lacey promised Sholler at least $80,000 for the phase I trial and followed through. Will later enrolled. He has participated in three of Sholler's clinical trials.

    Five years after Will was deemed incurable, he is, to his father's never-ending amazement, still alive. Will still has cancer—no treatment has managed to eradicate it—but for the moment his life is a lot like other children's. He's a hockey nut who started second grade last week, has two younger sisters, and he celebrated his 7th birthday in August.

    Lacey doesn't know why Will has experienced such good health for several years, something almost unheard of in relapsed neuroblastoma. “Why does a kid like Will come in, and for 3 years [Sholler] can keep him stable with an incredible quality of life?” Lacey wonders. Lacey would like to enroll his son in Sholler's drug-tumor matching study, but Will's tumor sits on his brachial plexus, a cluster of nerve fibers that stretch out from the spine. Biopsying it, Lacey and his wife have been told, could permanently damage the function of Will's arm. For now, they plan to wait.

  7. U.S. Science Education

    Two-Year Colleges Are Jumping Into the U.S. Research Pool

    1. Alison McCook*

    A growing number of community colleges hope to improve instruction and train a more diverse cadre of scientists by involving students in research.

    Political science.

    Chemistry professor David Brown poses at the White House with Southwestern College students, including Hector Mendoza Solano (second from right).


    Growing up in central Mexico, Héctor Mendoza Solano never imagined talking to a member of Congress in Washington, D.C. And the idea that the conversation would be about his research would have seemed unthinkable to his family, which couldn't afford to send him to university. Yet in April 2008, Solano found himself doing exactly that.

    Solano was attending the 2-year Southwestern College outside San Diego, California, and his research on glass materials with special optical properties was part of a “Posters on the Hill” event sponsored by the Council on Undergraduate Research (CUR), an organization formed in 1978 to foster undergraduate research. It was during that event that Solano—who crossed the border every day from Tijuana on an educational visa—learned he had been accepted into a chemical engineering program at the University of California, San Diego (UCSD). This June he received his bachelor's degree, and he plans to pursue a graduate degree in materials science after spending a few years in industry.

    Solano's story is more than an uplifting tale of personal achievement. It's also part of a burgeoning effort to train more scientists by offering research experiences to a huge but previously neglected pool of talent. The pool consists of students like Solano who attend so-called community colleges. There are more than 12 million such students—nearly half of all U.S. undergraduates—and each year about 25% transfer to 4-year schools. Although CUR was formed in 1978, community colleges are relative newcomers to its ranks and make up only a dozen of its 600 institutional members.

    Last year, CUR began running workshops to help them integrate research into their curricula. This month, the National Science Foundation (NSF), which has funded those workshops, gave $3.35 million to biologist James Hewlett of Finger Lakes Community College in Canandaigua, New York, to help 16 colleges that participated in those workshops take the next step by developing original research programs.

    They face significant challenges. Southwestern College chemistry professor David Brown spends up to 18 hours a week giving lectures and running instructional labs, a typical course load for community college faculty members. In addition to lacking assistants, Brown must send students, who may be gone after 2 years, into the field before they have acquired a solid academic foundation.

    Brown takes his students on field trips to UCSD because Southwestern College can't afford subscriptions to most specialized journals in his field. And community college faculty members can't really compete against larger, better-funded research teams. In 2003, Hewlett's students began a project to analyze DNA from local red-tailed hawks for segments that distinguished males from females. He's only now writing up the results, which essentially duplicate what two other groups have already published.

    But Hewlett isn't trying to finish first. His goal is to get students excited about science, with the added benefit of being energized himself. “The faculty who get involved in research become better teachers,” he says.

    Learning by doing

    Hewlett and his co-investigators on the NSF grant are still filling out their roster. They are looking for institutions like Portland Community College in Oregon, the state's largest educational institution with nearly 100,000 students. The NSF project will enhance what some faculty members are already doing, says biologist Josephine Pino, by offering salary support for faculty to reduce their teaching load and additional training to help them develop research-based courses. Such a course, Pino says, might make use of the college's 40 hectares of natural forest, stream, and wetlands.

    Hewlett came to Finger Lakes in 1998 already convinced of the value of student research. “Even though I was expected to teach all these classes, at some point I was determined to figure out how to do research, which was what got me excited about science in the first place,” says Hewlett, who holds a master's degree in molecular ecology from nearby University of Rochester.

    His students don't just get excited when they do research, he says. They also learn better. “My students knew more about DNA from this [red-tailed hawk] project than any students I'd ever had in the past doing a general biology course,” he says. A second course that lets them analyze how tart cherry juice reduces cell damage gives students a glimpse into the world of cell structure and function along with basic chemistry and the principles of experimental design. A course examining the bleaching of a tropical reef ecosystem teaches students about population ecology and the regulation of gene expression.

    To be sure, it takes more time to guide 30 students through original research projects than to simply lecture them. So Hewlett has obtained grants to ease his teaching load, purchase equipment, and support a full-time technician. He's also helped to document the indirect benefits of research for the college, including the fact that an initial $24,000 investment in research generated $1.2 million in new industry collaborations, external funding, or free materials and supplies.

    When M. Gita Bangera accepted a faculty position at Bellevue College outside Seattle, Washington, she assumed her students at the 2-year college would be perfectly capable of doing research. “I'd always done research, so I didn't see why that wouldn't be necessary,” says Bangera, a molecular biologist who had completed multiple postdoctoral fellowships and worked as a senior scientist at a micro-array manufacturer. In 2007, Bangera and fellow Bellevue faculty members Jim Ellinger and Chris Shelley obtained a $500,000 NSF grant to start a graduate school–type project called ComGen, in which students conduct research, write up a poster or manuscript, participate in journal clubs, and serve as teaching assistants in introductory biology courses.

    The project teaches students how to maintain a lab notebook, isolate plasmid DNA, and run PCR while they sequence the genome of Pseudomonas fluorescens L5.1-96, a bacterium that fights off a fungus that attacks wheat. They also analyze original research articles. “The first time they do it,” she says, “they're terrified.” One common complaint is about having to look up every second word in the article to understand it. “And I say, ‘Yeah, you're going to have to do that.’ They learn to ask and answer questions, which is really what research is about.”

    Brown believes that the data his students collect are just as good as those generated at a major research university, even if the scope of the project may be more limited. And his students are more likely to work directly with a faculty member, he adds. “I think my students are probably getting at least as good if not a better experience as someone who's doing research at a [major] university,” he says.

    However, that experience comes with a distinctly community college flavor. At Delaware Technical & Community College in Stanton, biochemist Virginia Balke asks students to help with projects that vary from the characterization of soil microbes to the population genetics of local bat and fox populations. But they must stop work at 5:15 p.m. on Fridays and 2:30 p.m. on Saturdays, which is when the only building on campus closes. And the building doesn't reopen until Monday.

    Learning by doing.

    Finger Lakes Community College students work in the field and in the lab.


    “You have to really plan the timing of your experiments well,” Balke says. She also must prepare for contingencies: The school has no backup generator, so when the electricity went out last summer, Balke brought in dry ice to cool a refrigerator full of fox feces, a carcass, and some body parts.

    The broader impact

    Hewlett and his colleagues hope to inspire their students not only to stay in school but also to view science as a viable career option. Their success could mean new opportunities for groups traditionally underrepresented in the sciences and engineering.

    Some 30% are African American or Hispanic, and nearly 60% are women. There is preliminary evidence that community college students who engage in research are more likely to stay in science, transfer to a 4-year school, and pursue a higher degree. Those outcomes suggest to Hewlett and others that research experiences can be a valuable tool to broaden the U.S. scientific talent pool. “If we believe that an early research experience is something that attracts under-represented groups and retains them in science careers, it makes sense to support that,” says V. Celeste Carter, program director within NSF's division of undergraduate education in Arlington, Virginia.

    Hewlett has found that students taking courses that include original research are more likely to transfer to 4-year science programs (62% compared with 51% before 2007). The percentage of those applying to graduate school in science after obtaining bachelor's degrees from another school has grown to 10%, from 4%. Also, 36% of his students said they planned to transfer to a 4-year science program because of their research experience, and another 36% said their research provided them with an internship or job they wouldn't have otherwise.

    For Brown, the value of community college research is captured in Solano's reaction to his 2007 trip to UCSD to photocopy journal articles. Brown recalls Solano telling himself during the visit, “‘Okay, this is the place where I want to be.’”

    Solano believes that his research experience was the key to being accepted by UCSD as a transfer student. That research is still opening doors as he seeks a job with industry. “I got a call the other day for a research position,” he says. “And they were really interested in my research experience with Dr. Brown.”

    From Community College to NIH


    R. Douglas Fields

    Developmental neurobiologist and section chief

    National Institute of Child Health and Human Development

    Started at a community college in Cupertino, California, and wouldn't have it any other way.

    “Without question, I owe my career to the outstanding educational program at De Anza College.” Doing research there was a great introduction to the profession, he adds: “You can't learn science any other way.”


    Elaine Ostrander

    Chief of Cancer Genetics Branch

    National Human Genome Research Institute

    Began college at Yakima Valley Community College in Washington state because of its lower tuition. Working part-time to pay for classes didn't leave her any time to do research, she says, but community college students should have that option.

    “I absolutely wish I had had that opportunity.”

    • * Alison McCook is a freelance writer based in Philadelphia, Pennsylvania.

Log in to view full text

Via your Institution

Log in through your institution

Log in through your institution