'Marsquake' Hunter Begins to Probe Planet's Innards

NEWS IN FOCUS

claimed experiments for an article in MIT In an interview with the Associated Press, online condemning He’s claims. “Directly Technology Review. “The data I reviewed are He said the goal of the work was not to prevent jumping into human experiments can only consistent with the fact that the editing has, in transmission from the parents, but to offer be described as crazy,” the statement reads. fact, taken place,” says Urnov, who is based at couples affected by HIV a chance to have a The scientists call on Chinese authorities to the Altius Institute for Biomedical Sciences in child that might be protected from a similar release the findings of any investigation to the Seattle, Washington. But he adds that the only fate. But years of research is needed to show public. way to tell whether the children’s genomes have that meddling with the genome of an embryo “This is a huge blow to the international been edited is to independently test their DNA. is not going to cause harm, says Joyce Harper, reputation and the development of Chinese Urnov takes issue with the decision to edit who studies women’s and reproductive health science, especially in the field of biomedical an embryo’s genome to prevent HIV infection. at University College London. Legislation and research,” the statement says. “It is extremely He is also using genome-editing tools to target public discussion should also occur before unfair to the large majority of diligent and con- the CCR5 gene, but his studies are in people genome editing is scientious scientists in China who are pursuing with HIV, not embryos. He says that there are used in embryos “This is a huge research and innovation while strictly adher- “safe and effective ways” to use genetics to pro- destined for implan- blow to the ing to ethical limits.” tect people from HIV that do not involve edit- tation. international Nature tried to contact He but did not ing an embryo’s genes. Southern Univer- reputation and receive a response before its deadline. In his Paula Cannon, who studies HIV at the Uni- sity of Science and the development video, He says he supports the use of genome versity of Southern California in Los Angeles, Technology said of Chinese editing in embryos only in cases that relate to also questions He’s decision to target that gene in a statement on science.” disease. “I understand my work will be contro- in embryos. She says that some strains of HIV 26 November that versial, but I believe families need this technol- don’t even use this protein to enter cells, they it was unaware of He’s experiments, that the ogy and I am willing to take the criticism for use another protein called CXCR4. Even people work was not performed at the university and them,” he says. who are naturally CCR5-negative are not com- that He has been on leave since February. The News of the experiment came a day before pletely resistant to HIV, Cannon adds, because university says its researchers must abide by researchers in the field gathered in Hong Kong they could be infected by a CXCR4 strain. national laws and regulations, and respect for a major international meeting on genome She also says it makes no sense that He international academic ethics and academic editing, running from 27 to 29 November. recruited families with an HIV-positive father, standards. It will set up an independent com- Even before the news of He’s work emerged, as was the case with the twins, because there mittee to investigate the matter. many in the field thought it was inevitable that is no real risk of transmission to the children. Making gene-edited babies goes against regu- someone would use genome-editing tools to “This experiment exposes healthy normal lations released by China’s health and science make changes to human embryos for implan- children to risks of gene editing for no real ministries in 2003, but it is not clear whether tation into women, and had been pushing for necessary benefit,” says Julian Savulescu, direc- there are penalties for those who break the rules. an international consensus on how genome tor of the Oxford Uehiro Centre for Practical More than 100 Chinese biomedical research- editing to modify eggs, sperm or embryos Ethics at the University of Oxford, UK. ers posted a strongly worded statement should proceed. ■

PLANETARY SCIENCE ‘Marsquake’ hunter begins to probe planet’s innards Joint US-French-German mission will monitor seismic activity on Mars.

BY ALEXANDRA WITZE (JPL) in Pasadena, California. On Monday, which needs a safe, geologically stable place to just before 11:53 a.m. local time, the space- do its work. arthlings are about to hear Mars’s heart- craft entered the Martian atmosphere at nearly The first photo that InSight sent from the beat. 20,000 kilometres per hour. surface of Mars showed a flat, relatively rock- On 26 November, NASA’s InSight As it neared Mars’s surface, the spacecraft free landscape stretching to the horizon, with Emission touched down near the Martian demonstrated a new way to communicate with the foreground speckled with dust from the equator and embarked on the first mission its controllers on Earth, 146 million kilometres landing. dedicated to listening for seismic energy away. Two ‘cubesats’, each the size of a briefcase, “It’s happy. The lander is not complain- rippling through the red planet. relayed information from InSight to Earth in ing,” said Rob Manning, chief engineer at JPL, Any ‘marsquakes’ InSight detects could yield close to real time. The experiment suggests that shortly after InSight touched down. clues about the planet’s mysterious interior, miniature satellites like these could allow faster including how it is separated into a core, man- communication with probes in deep space. LISTENING IN tle and crust. Whatever scientists learn about InSight landed at Elysium Planitia, a broad, Mission scientists will use the lander’s cam- Mars’s innards could help to illuminate how flat region just north of the Martian equator. It era to scout the ground for the smoothest and our own planet evolved billions of years ago. is one of the most boring places on the planet, most level area to deploy its French-built seis- InSight had been cruising through space says Bruce Banerdt, a planetary scientist at JPL mometer (see ‘Ear to the ground’). InSight’s since its launch in May, tracked by mission and the US$994-million mission’s principal robotic arm will pluck the instrument off its control at NASA’s Jet Propulsion Laboratory investigator. That’s an advantage for InSight, back and place it on the ground, then put a

IN FOCUS NEWS

result of jostling by the wind or other sources. EAR TO THE GROUND On the third day after landing, project sci- NASA’s Mars InSight lander will gather data on seismic activity to help scientists better understand the red entists will switch on an instrument to track planet’s mysterious interior. changes in the magnetic field, which will help

NASA/JPL-CALTECH The 1.8-metre robotic arm will Two solar panels will supply them to identify sources of noise that aren’t place a seismometer and heat power to the lander and its quakes, says Catherine Johnson, a geophysi- probe onto Mars’s surface. instruments. cist at the University of British Columbia in Vancouver, Canada. InSight won’t deploy its German-built heat- flow probe until January. Over the course of several weeks, the instrument will drill five metres into the Martian surface, deeper than The lander’s anything achieved before. Scientists will track seismometer will listen changes in temperature as small as a few hun- for tremors known as dredths of a degree. That will tell them how marsquakes. much heat is leaving Mars, and how many A heat probe will dig down 5 metres to measure heat-producing radioactive elements are temperature change over packed inside it. depth and time. InSight is meant to work for a little more than one Martian year, equivalent to almost two Earth years. It should measure 50–100 dome-shaped wind shield over it. The whole its sensitivity allows it to detect movement as marsquakes during that period, says Banerdt. process is expected to take several days. small as the width of an atom. The big chal- The longer it survives, the more it will be able The seismometer includes three ground- lenge will be determining which movements to detect — and the more researchers will be motion sensors nested inside a vacuum, and are caused by marsquakes and which are the able to deduce about Mars’s internal structure. ■

PUBLISHING The age of AI peer reviews Automated software can help review papers, but the decision-making stays with humans.

BY DOUGLAS HEAVEN phrases in what they have actually written,” he concepts. He says that this kind of tool will be says, “instead of just taking what they’ve come useful beyond peer review, for tasks such as ost researchers have good reason up with five minutes before submission.” writing grant applications or literature reviews. to grumble about peer review: it is UNSILO identifies which of these key phrases Many platforms, including ScholarOne, time-consuming and error-prone, are most likely to be claims or findings, giving already have automatic plagiarism checkers. Mand the workload is unevenly spread, with editors an at-a-glance summary of the results. And services including Penelope.ai examine just 20% of scientists taking on most reviews. It also highlights whether the claims are simi- whether the references and the structure of a Now peer review by artificial intelligence lar to those from previous papers, which could manuscript meet a journal’s requirements. Some (AI) is promising to improve the process, be used to detect plagiarism or simply to place can flag up issues with the quality of a study, boost the quality of published papers — and the manuscript in context with related work in too. The tool statcheck, developed by Michèle save reviewers time. A handful of academic the wider literature. “The tool’s not making a Nuijten, a methodologist at Tilburg University publishers are piloting AI tools to do anything decision,” says Chris- in the Netherlands, and her colleagues, assesses from selecting reviewers to checking statistics tensen. “It’s just say- “It doesn’t the consistency of authors’ statistics reporting, and summarizing a paper’s findings. ing: ‘Here are some replace editorial focusing on P values. The journal Psychological In June, software called StatReviewer, things that stand out judgement but, Science runs all its papers through the tool, and which checks that statistics and methods in when comparing by God, it makes Nuijten says that other publishers are keen to manuscripts are sound, was adopted by Aries this manuscript with it easier.” integrate it into their review processes. Systems, a peer-review management system everything that’s been When Nuijten’s team analysed papers owned by Amsterdam-based publishing giant published before. You be the judge.’” published in psychology journals, they found Elsevier. And ScholarOne, a peer-review plat- “It doesn’t replace editorial judgement but, by that roughly 50% contained at least one statis- form used by many journals, is teaming up God, it makes it easier,” says David Worlock, a tical inconsistency (M. B. Nuijten et al. Behav. with UNSILO of Aarhus, Denmark, which UK-based publishing consultant who saw the Res. Meth. 48, 1205–1226; 2016). In one in uses natural language processing and machine UNSILO demonstration at the Frankfurt Book eight papers, the error was serious enough that learning to analyse manuscripts. Fair in Germany last month. it could have changed the statistical signifi- UNSILO uses semantic analysis of the Worlock notes that there are several similar cance of a published result. “That’s worrisome,” manuscript text to extract what it identifies as tools emerging. He is on the board of Wizdom.ai she says. She’s not surprised that reviewers miss the main statements. This gives a better over- in London, a start-up owned by publishers Tay- such mistakes, however. “Not everyone has view of a paper than the keywords typically sub- lor & Francis, which is developing software time to go over all the numbers. You focus on mitted by authors, says Neil Christensen, sales that can mine paper databases and extract the main findings or the general story.” director at UNSILO. “We find the important connections between different disciplines and For now, statcheck is limited to analysing