News this Week

Science  27 Feb 2015:
Vol. 347, Issue 6225, pp. 930

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

  1. This week's section

    Science is beautiful

    This false-color x-ray of a snapping turtle revealed about 30 eggs hidden inside.


    A snapping turtle's hidden eggs, a forest of neurons, and the seahorse-shaped hippocampus were among the colorful, complex muses for the winning entries in the 2015 Vizzies, or Visualization Challenge, sponsored by the National Science Foundation and Popular Science. Each category—photography, video, illustration, posters & graphics, and games & apps—includes an Experts' Choice prize of $2500 and a People's Choice prize of $1000. The Experts' Choice winner for photography, shown here, was a false-color x-ray image of a snapping turtle that revealed several dozen eggs inside her. In the posters & graphics category, the Experts' Choice prize went to a creative depiction of the brain's hippocampus, responsible for acquiring some memories; real hippocampal neurons were digitally added to a pair of seahorses. Both the Experts' Choice and People's Choice prizes in illustration went to “Neuroforest,” an image taken from a graphic novel depicting the adventures of a man who fell into a brain and wanders through a dense forest network of brain cells and neurons. A full list of winners and more information on the competition is available at

    Shall we play a game?

    An AI system has mastered Space Invaders and 48 other classic Atari games.


    Those Space Invaders may have finally met their match. A team of scientists has developed an artificial intelligence system that has taught itself to play dozens of classic computer games—and in many cases, it plays them better than professional gamers, the researchers note this week in Nature. Led by Volodymyr Mnih of Google DeepMind in London, the team trained their new artificial “agent” to master 49 different Atari 2600 games—which requires adapting to a variety of tasks—solely from its study of data, rather than from any programmed behavior. The agent, called a deep Q-network, combines reinforcement learning techniques with a biologically inspired artificial neural network known as a deep neural network. To “learn” how to beat the games (and human games testers), the deep Q-network examined large data sets and learned what would increase its score.

    “We have so much great data … this data is the citizens' data.”

    New U.S. Chief Data Scientist DJ Patil, in a 20 February “memo to the American people” outlining how he will advise the White House on investing in data science and research.

    Around the world


    E.U.'s first stem cell therapy

    Europe's first commercial stem cell product, developed to reverse blindness caused by burns and chemical damage to the eye's surface, was authorized by the European Commission on 19 February. The therapy, called Holoclar, uses stem cells from the healthy part of a patient's limbus (the border between the cornea and the white of the eye). The cells are lab-cultured and then implanted into the damaged eye. Holoclar was first developed by the University of Modena and Reggio Emilia in Italy and will be commercialized by Chiesi Farmaceutici. The commission's final approval followed the European Medicines Agency's green light for the therapy in December 2014.

    Geneva, Switzerland

    New rapid Ebola test

    The World Health Organization (WHO) has approved the first rapid diagnostic test for Ebola. Unlike current PCR-based tests, which require a blood sample taken by needle, secure transport of the blood to a properly equipped laboratory with trained staff, and at least several hours to return results, the new test needs no electricity, requires just a few drops of blood from a finger prick, and can return results in 15 minutes. Produced by Broomfield, Colorado–based Corgenix, the test uses antibodies to identify a specific Ebola virus protein. WHO notes that it correctly identifies 92% of infected people and 85% of uninfected ones. It will cost about $15, although discounts will be available for bulk purchases and suppliers for use in Africa.

    Homalin, Myanmar

    Resistant malaria nears India

    Malaria parasites that are resistant to the most powerful antimalaria drug may have spread to the border of Myanmar and India, a potential disaster for global malaria control. A genetic study of malaria parasites across Myanmar, published online on 19 February in The Lancet Infectious Diseases, found widespread mutations in a gene that plays a key role in resistance to artemisinin, the best drug available to cure the disease. Follow-up studies are needed to confirm that the mutations correspond to clinical resistance in patients, but the pattern is worrisome, the authors say. Resistant parasites appeared a decade ago in Cambodia and Thailand, and experts have been warning that they would spread. The new study suggests they may have spread farther and faster than experts had realized—or feared.


    Nature offers double-blind reviews

    Nature Publishing Group (NPG) announced last week that most of its publications, including flagship journal Nature, will offer authors the option of double-blind peer review, in which both submitters and referees remain anonymous. Traditionally, scientific journals adhere to a single-blind system in which authors don't know the identity of reviewers, but that has led to concerns about antiauthor biases. NPG began testing the double-blind system with Nature Climate Change and Nature Geoscience in May 2013, a trial that “gave us plenty to be confident about,” says Véronique Kiermer, director of author and reviewer services at NPG. But, she says, the option of complete anonymity will be “an ongoing experiment.” Meanwhile, editors at other journals, including Science, haven't made the jump, in part out of concern that reviewers may still be able to guess the identity of authors, particularly in small fields.


    Court overrules misconduct panel

    A Danish court has overturned a committee's verdict of scientific misconduct and awarded 400,000 Danish kroner ($61,000) in legal expenses to a physiologist at the University of Copenhagen. In December 2013, the Danish Committees on Scientific Dishonesty found Bente Klarlund Pedersen guilty of “gross negligence” for failing to detect images that had been manipulated by a co-author and found that her reuse of biopsy samples in multiple papers constituted “unclear construction of data.” Klarlund challenged the verdict in court, arguing that using biopsies for multiple studies was common practice and that failing to detect fraud committed by a co-author was a mistake but not gross negligence. On 18 February the court ruled in her favor on both counts, stating that there is a “crucial difference” between deliberately manipulating images and failing to detect manipulations by collaborators.

    IPCC head resigns

    Rajendra Pachauri, chair of the Intergovernmental Panel on Climate Change (IPCC) since 2002, stepped down on 24 February amid allegations of sexual harassment by a female colleague at The Energy and Resources Institute, a New Delhi–based think tank led by Pachauri. In a letter addressed to U.N. Secretary-General Ban Ki-moon, Pachauri said that “[t]he IPCC needs strong leadership and dedication of time … which, under the current circumstances, I may be unable to provide.” The staffer's complaint, now under investigation by local police, states that Pachauri sent her messages with sexual undertones. Pachauri's lawyers have countered that his computer had been hacked; Pachauri himself was unavailable for comment. Vice Chair Ismail El Gizouli will be acting IPCC chair until elections are held for the position in October.


    Eating peanuts cuts allergy risk

    Peanuts consumption may prevent allergies.


    In news that electrified the allergy world this week, a group of scientists reported that eating peanuts slashes the chance of an allergy, at least in children at high risk of developing one. The trial, published in The New England Journal of Medicine, is the largest and longest running of its kind, enrolling 640 babies and following them until they were 5 years old. It found that 17% of children who avoided peanuts became allergic, compared with 3% of those who ate modest amounts several times a week. The authors argued that many babies and toddlers who avoid peanut develop allergy because they are still exposed through the skin and that ingesting peanut products helps the body learn to tolerate them. The results are likely to catapult a long-standing theory into mainstream medicine.