Opening up the Scientific Literature Meet the New EMBO Young
Total Page:16
File Type:pdf, Size:1020Kb
WINTER 2017 ISSUE 37 28 researchers selected Meet the new EMBO Young Investigators PAGES 4 – 5 Award for SourceData Open Access Opening up the scientific Going for gold for a literature sustainable future PAGE 3 PAGE 11 Transparency at EMBO Press Recognition of excellence Celebrating 10 years ERC Responsible and accountable publishing Newest EMBO Members welcomed Frank Gannon recounts how it came into being at annual meeting PAGE 6 PAGE 7 PAGE 10 www.embo.org Table of contents EMBO NEWS SCIENCE POLICY © Marietta Schupp, EMBL Photolab Marietta Schupp, © SourceData wins publishing Open Access: why green is not the new innovation award Page 3 gold Page 11 Editorial Welcoming 28 new Young ith the announcement that all publicly funded research in Investigators Pages 4 – 5 WEurope must be freely acces- sible from 2020 onwards, the European EMBO Press brings more transparency Union firmly anchored Open Access in the to publishing Page 6 future of scientific publishing. It’s an ambi- tious target, and in the quest for solutions, we mustn’t overlook the need for publish- Women in Science Awards ing models that are sustainable in the long presented Page 6 term (p 11). SCIENCE STORY At the same time, the ‘when’ and ‘how’ to make scientific papers openly available Non-coding RNAs – where are not the only questions we must think about. Making an article available for free next? Pages 12 – 13 on the internet does not necessarily mean that interested readers, whether they are scientists or not, are able to find and bene- EMBO community fit from the publication. Back in 2009, the 50 millionth scientific Updates from across Europe article was published. With an estimated growth rate of between 8 and 9%, we will Pages 14 – 16 EMBO NEWS be approaching double that figure in the next two years. Adding to that the fact that the figures and data, which lie at the heart Newest EMBO Members meet in Awards and publications of a scientific publication, are not easy to Heidelberg Page 7 search, it is becoming increasingly difficult Achievements and papers by members to identify relevant work and keep up with EMBO at Basel Life: a snapshot Page 8 the latest studies in a field. of the EMBO community Page 17 EMBO’s SourceData platform offers a solution by turning figures into search- EMBO | EuropaBio Fellowships open able entities (p 3). By creating machine- for application Page 9 EMBO events readable descriptions of scientific figures, SourceData makes scientific data more Strengthening ties with the Taiwanese Upcoming courses, workshops and discoverable. A proof-of concept showed that the approach works. The next step will Ministry of Science and Technology conferences Page 17 be to work with the community to embed Page 9 SourceData more widely to make sure we really can open up the scientific literature. Frank Gannon reflects on the birth of Fresh from EMBO press the ERC Page 10 Some of our latest publications at a Maria Leptin Director, EMBO glance Page 19 2 EMBOencounters | Winter 2017 | [email protected] ©2017 EMBO EMBO NEWS Opening up the scientific literature How SourceData aims to combine scientific expertise with machine learning n the biological sciences, most of the data Generating the metadata from figures that are supervision to ensure the quality of the informa- produced are published in the form of figures. necessary for SourceData’s capabilities means tion extracted, explains SourceData project lead IHowever, the search tools used to find identifying the biological entities – the proteins, Thomas Lemberger. “The important task of iden- published papers are usually limited to keyword- molecules or genes – referenced in each figure tifying discrete, named entities from text input based text searches. and legend, as well as their relationships. is one that can be addressed by automated text- EMBO’s SourceData platform generates a Eva Benito-Garagorri, who recently joined mining technologies.” description of these figures, and the data and SourceData as Scientific Coordinator, explains To explore how text-mining can be used as relationships they contain, in a machine-reada- further: “Interpreting this information is rela- part of the curation workflow, SourceData joined the 6th BioCreAtIvE (Critical Assessment of Information Extraction systems in Biology) chal- lenge. As part of the challenge text-miners are SourceData wins ALPSP award invited to use the SourceData dataset to train and assess algorithms that match entities found in SourceData received recognition from the which brought together many of the major figure legends with their identifiers from stand- scientific publishing community when it was publishing houses. Both SourceData and Publons ardised biological taxonomies such as Uniprot or awarded the ALPSP Award for Innovation in (publons.com) received awards for innovation ChEBI, and then output the results in a suitable Publishing 2017. The annual award from the and were praised by the chair of judges, David format for SourceData. Association of Learned and Professional Society Sommer, who said: “Our winners are the result of Publishers (ALPSP) recognises outstanding inno- successful collaboration and partnership work- Opening new avenues with artificial intelligence vators in scholarly and professional publishing. ing within the industry to make a positive differ- The potential for automation in the SourceData SourceData was presented with the award on 15 ence to the academic community for the good of curation workflow does not end with entity recog- September 2017 at a ceremony in Amsterdam all.” nition. Thomas Lemberger explains: “We know that machine learning is making rapid progress in being able to extract meaning from natural language. Such methods are used by Twitter to identify sentiment, and by Google Translate to interpret the meaning of phrases. In our case, the task is to identify the nature of the relationship between entities – which one is influencing the other?” While this methodology remains in develop- ment, Lemberger is looking towards new para- digms whereby artificial intelligence algorithms improve themselves by learning from corrections made by expert curators and authors during publication. Successful implementation would make SourceData one of the first examples in publishing of coupling the expertise of scien- tists with the capabilities of machine learning models. This would open new avenues in scien- tific publishing, extending SourceData’s mission to the extraction and organization of structured Thomas Lemberger accepts the ALPSP award 2017. scientific knowledge. © The Association of Learned & Professional Society Publishers & Association of Learned The © More information: sourcedata.embo.org ble format. As a result, articles can be searched tively easy for humans with the appropriate Liechti R, George N, Götz L, El-Gebali S, Chasapi A, based on the data shown in their figures, and scientific background, so figures are currently Crespo I, Xenarios I & Lemberger T (2017) SourceData figures that display related results can be linked processed for SourceData by professional bio- - a semantic platform for curating and searching across papers. The method for this has now been curators.” However, developments in machine figures, Nature Methods 14: 1021-1022. Doi: published in a paper authored by EMBO project learning could mean that the future holds a faster 10.1038/nmeth.4471 team members and their collaborators at the and easier way to extract the information from Editorial: The search for data, Nature Methods 14: Swiss Institute of Bioinformatics (Liechti et al., figure legends. 1017 Nature Methods, 2017). The SourceData approach has the potential Automation for faster curation to accelerate scientific discovery and “could The SourceData platform remains in active devel- fundamentally shift how scientists interact with opment with the goal of establishing an open and the collective body of knowledge and open up effective standard for the discovery and re-use of all data to synthesis, reassessment and reuse”, as scientific data. One of the next steps in develop- highlighted in a recent editorial (Nature Methods, ing the platform is the acceleration of the figure November 2017). curation workflow while keeping expert-based ©2017 EMBO EMBOencounters | Winter 2017 | [email protected] 3 EMBO NEWS Welcoming the new EMBO Young Investigators 28 group leaders from 11 countries join the EMBO community Filipe Cabreiro Bacteria: macromolecular machines for nutrition and 28 young researchers were elected drugs as EMBO Young Investigators London, GB this year. They join a network of 47 current and 417 past Young Investigators who represent some of the best up-and-coming group leaders in the life sciences in Europe and beyond. “It is a pleasure to welcome these Axel Innis Nicola Iovino Manuel Irimia Ribosome inhibition Epigenetic regulation of Transcriptomics of outstanding scientists to the by nascent or fertilization vertebrate development antimicrobial peptides Freiburg, DE and evolution EMBO community,” says EMBO Bordeaux, FR Barcelona, ES Director Maria Leptin. “Between them, they carry out some very promising life science research that Europe has to offer, and we look forward to supporting them in their professional and scientific endeavours.” Daniel Messerschmidt Patrick Müller Elizabeth Murchison Epigenetic reprogramming Quantitative analysis, Genetics and evolution The new Young Investigators are and its implications in control,and engineering of of transmissible cancers based in 11 different