Message from ISCB Live Coverage of Intelligent Systems for Molecular Biology/European Conference on Computational Biology (ISMB/ECCB) 2009

Allyson L. Lister1,2*, Ruchira S. Datta3, Oliver Hofmann4, Roland Krause5, Michael Kuhn6, Bettina Roth7, Reinhard Schneider8 1 Centre for Integrated Systems Biology of Ageing and Nutrition, Institute for Ageing and Health, Newcastle University, Campus for Ageing and Vitality, Newcastle upon Tyne, United Kingdom, 2 School of Computing Science, Newcastle University, Newcastle upon Tyne, United Kingdom, 3 QB3 Institute, University of California, Berkeley, Berkeley, California, United States of America, 4 Harvard School of Public Health, Department of Biostatistics, Boston, Massachusetts, United States of America, 5 Department of Computer Science, Free University Berlin and Department of Computational Molecular Biology, Max Planck Institute for Molecular Genetics, Berlin, Germany, 6 Biotec, Technische Universita¨t Dresden, Dresden, Germany, 7 International Society for Computational Biology, La Jolla, California, United States of America, 8 Computational and Structural Biology Unit, European Molecular Biology Laboratory (EMBL), Heidelberg, Germany

ing, shortly before the start of the first gauer’s keynote received the most attention, presentation. The feeds are accessible on with more than 230 comments, equating to a the conference pages in the appropriate comment every 15 seconds. The mean program sections, with the most recent number of comments over all blogged talks comments visible on the main ISCB portal inthemainsessionwas22.Therewere40 entry page. They can also be found on active bloggers, amounting to roughly 3% of FriendFeed by searching for the author conference attendance. There were more Introduction name, the title of the talk or the talk identifier than 140 subscriptions to the FriendFeed as given in the program. The live blogging room, comprising both remote and local The International Conference on Intel- event was advertised in various ways: attendees. ligent Systems for Molecular Biology mailings to all attendees and ISCB members, Live blogging became so popular at (ISMB) 2008 conference in Toronto was announcements in the printed ISCB news- ISMB/ECCB 2009 that, by the end of the probably the first life science conference letter (http://www.iscb.org/images/stories/ first day of the main conference, the security that saw a meeting report based on newsletter/newsletter12-1/index.htm), links procedures at FriendFeed automatically microblogging activity. This activity started from the portal and conference Web sites, blocked the IP address of the conference several days before the conference through advance notification of journalists, and network because of abnormally high activity discussions on the FriendFeed (http:// advertisements in The Life Scientists room (http://ff.im/4zHLB). Within about an www.friendfeed.com) platform and the on FriendFeed (http://friendfeed.com/the- hour, FriendFeed responded to feedback creation of an unofficial room dedicated life-scientists). Other platforms such as Twit- from conference attendees and reversed the to the conference. The use of this room led ter (http://www.twitter.com) and personal block. To many local and remote attendees, both to a physical meetup and a meeting blogs were used at the conference. Because it this signalled the beginning of live blogging report [1]. The main focus of the bloggers is difficult to retrieve Twitter statistics and as a mainstream conference tool (http:// was their shared interest in science com- this platform was not used extensively during blog.openwetware.org/scienceintheopen/2009/ munication. Unsubstantiated opinions and the ISMB/ECCB 2009, we focus on the 06/29/conferences-as-spam-liveblogging- off-topic comments were therefore very reporting activities in the FriendFeed room. science-hits-the-mainstream/). limited and the general feeling was that it More than half of the presentations in the Keynotes was a positive experience which brought main session (56%, or 93 out of 165) had benefits for the broader conference audi- FriendFeed comments in the ISMB/ECCB The keynote talks ranged from abstract ence as well as for scientists following the 2009 room (see Figure 1), with peak blogging to applied topics, and from classical scientific program remotely. activity during the keynotes. Thomas Len- biology to cutting-edge computational As a result, the International Society for Computational Biology (ISCB), organizers of the Intelligent Systems for Molecular Citation: Lister AL, Datta RS, Hofmann O, Krause R, Kuhn M, et al. (2010) Live Coverage of Intelligent Systems Biology/European Conference on Com- for Molecular Biology/European Conference on Computational Biology (ISMB/ECCB) 2009. PLoS Comput Biol 6(1): e1000640. doi:10.1371/journal.pcbi.1000640 putational Biology (ISMB/ECCB) 2009 conference decided to actively support Editor: Olga Troyanskaya, Princeton University, United States of America future blogging efforts. The live blogging Published January 29, 2010 efforts described here can be seen as a Copyright: ß 2010 Lister et al. This is an open-access article distributed under the terms of the Creative model for future conferences, with the Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. organizers providing a tight link between Funding: RSD acknowledges funding under the Microbial Genome Sequencing Program of the National the FriendFeed ISMB/ECCB 2009 room Science Foundation (grant #0732065 to Kimmen Sjo¨lander). ALL is supported by the BBSRC/EPSRC funding for (https://friendfeed.com/ismbeccb2009) CISBAN [BB/C008200/1]. The funders had no role in study design, data collection and analysis, decision to and the conference Web site in the ISCB Web publish, or preparation of the manuscript. portal (http://www.iscb.org/ismbeccb2009/). Competing Interests: The authors have declared that no competing interests exist. Talk-specific feeds were created each morn- * E-mail: [email protected]

PLoS Computational Biology | www.ploscompbiol.org 1 January 2010 | Volume 6 | Issue 1 | e1000640 Figure 1. Summary of commented talks at ISMB/ECCB 2009. (A) Keynotes were the most commented talks at ISMB/ECCB 2009. Here, keynotes are listed in chronological order, showing that the number of comments per keynote was higher at the end of the conference than at the beginning. Nonetheless, the absolute number of comments does not necessarily reflect the quality of the coverage. (B) Besides the keynotes, in the main session the highlights and proceedings track received the most attention. However, the total number of covered talks in the SIGs was higher than in any other session. For simplicity, talks from the special sessions and SIGs are summarized across all special sessions and all SIGs, as the commenting method for those sessions was not as uniform as for the main sessions. More detailed statistics are available at the Web sites in Box 1. doi:10.1371/journal.pcbi.1000640.g001

PLoS Computational Biology | www.ploscompbiol.org 2 January 2010 | Volume 6 | Issue 1 | e1000640 Box 1. Session Statistics with protein networks to diagnose breast cancer metastasis. Describing the difficulty Main session: http://www.iscb.org/cms_addon/friendfeeds/ff_stats.php of extracting subnetworks or modules from the global ‘‘hairball’’ networks, Ideker even Special sessions; http://www.iscb.org/cms_addon/friendfeeds/ff_ssstats.php suggested that the hairball might be an essential feature of such networks. SIGs: http://www.iscb.org/cms_addon/friendfeeds/ff_sigsstats.php Both Thomas Lengauer and Mathias Uhlen discussed their research on applying Special sessions and main tracks combined: http://www.iscb.org/cms_addon/ friendfeeds/ff_statsincss.php computational biology to concrete endeav- ours. Lengauer, from the Max-Planck All three session types: http://www.iscb.org/cms_addon/friendfeeds/ff_statsincsssigs. Institute for Informatics and a founding php member of both the ISCB and the ECCB, provided an overview of his ongoing work in Chasing the AIDS Virus (http://ff.im/ biology. Pierre-Henri Gouyon, of the research focuses on Ecuadorian marsupial 4EvfT). HIV has a dynamic rate of Muse´um national d’Histoire naturelle, frogs (Gastrotheca), and how their develop- evolution, and with a turnover of more provided the first keynote of the confer- ment compares to other well-studied frog than 10 million virus particles per day per ence, and spoke on the abstract topic of families. Interestingly, the reproductive patient, a drug may be efficient against a information in biology (Information and and egg physiology (i.e., a comparatively wild type strain, but not against the mutants Biology, http://ff.im/4yWcl) and the diffi- larger size of egg), and longer time to the present in the infected population. Len- culty of defining information in a biolog- embryonic gastrulation stage of the mar- gauer models viral evolution to the resis- ical context. For Gouyon, biologists are supial frogs bear some resemblance to that tance using a tree structure, where every working with fuzzy and poorly defined of mammals. However, there is a definite branching represents several alternatives definitions of information. He describes lack of molecular data for exotic frogs, and for viral evolution. Using this technique, Richard Lewontin’s ‘‘triple helix’’ of such data could greatly help comparative Lengauer has discovered that mutations genes, environment, and epigenetics as approaches to frog developmental biology. can confer resistance to one drug but joint providers of information, and warns Daphne Koller and Trey Ideker described increase sensitivity to others. This result against simplifying or diminishing the recent advances in networks; while Koller could not have been found with mutation importance of epigenetic information. focused on her work with regulatory networks tables. Lengauer described how therapy Gouyon believes that it will take efforts (Individual Genetic Variation: From Networks to optimization via viral evolution modelling by biologists, computer scientists, and Mechanisms, http://ff.im/4BFeQ), Ideker dis- with their THEO prediction engine sup- physicists working together to define cussed the possibilities of using networks for ports personalized medicine in HIV/AIDS information in a way suitable for biology. disease diagnosis. Koller, from Stanford treatment, and helps the doctor make the Tomaso Poggio from the Massachusetts University, discussed how, by creating a difficult decision as to which combinations Institute of Technology gave another keynote framework for modelling gene regulation, of drugs to give to AIDS patients. THEO presentation, Computational Neuroscience: Models diverse regulatory mechanisms can be un- (THErapy Optimizer) has a therapy classi- of the Visual System (http://ff.im/4yVNH). covered and the effect of gene regulation on fication error rate of less than 15%, as Vision is more than simple categorization: it phenotype can be understood. Koller exploits opposed to 24% for standard methods. is also image understanding, inference, and modularity in regulatory networks, which Mathias Uhlen, of the Royal Institute of parsing. Poggio has built quantitative models provides increased statistical power, to predict Technology in Sweden and inventor of of the ventral stream for object recognition, expression of an entire module as well as pyrosequencing, believes the 21st century is modelling only the feed-forward connections, uncover coupled regulator programs. For the century of medicine, and that systems and is extending the models to videos and example, the transformation of follicular biology will play a strong role over the next sequences of images. Poggio has tested his lymphoma to diffuse large B cell lymphoma 10 years (A global view on protein expression models by comparing mouse behaviour, occurs in 40%–60% of patients, and though based on the Human Protein Atlas, http://ff.im/ which in turn can be used as a model system gene-based analysis was inconclusive, a 4Hk76). As proteins are primary drug for psychiatric diseases. When studying short- regulatory network model performed well, targets and there is a pressing need for duration mouse behaviours, Poggio’s system and may have therapeutic implications using antibodies as protein probes, Uhlen devel- wasalmostasgoodashumanobservers: connectivity map analysis. oped the Human Antibody Initiative (HAI), while classifications by different humans The ISCB Overton Prize Lecture for Antibodypedia (http://www.antibodype- agreed 72% of the time, there was 71% 2009 was given by Trey Ideker of Univer- dia.org), and the Human Proteome Re- agreement between humans and the system. sity of California, San Diego. Ideker source (HPR, http://www.proteinatlas. Further, Poggio’s model can be used to infer described systems biology as a research org/). The HAI analyzed the success rate the mouse strain from behaviours with about area for jacks-of-all-trades, rather than of more than 500 commercially available 50% accuracy in 10 minutes of video. masters of one (New Challenges and Opportu- antibodies from 51 commercial companies, Discussing her work on frog develop- nities in Network Biology, http://ff.im/ and discovered a wide variety of successes mental biology and how it could be aided 4BERS). His main focus was on the and failures, which can be accessed through by additional computational techniques, emerging field of network-based disease Antibodypedia. The HPR is a multidisci- Eugenia del Pino from the Pontificia diagnosis. Ideker is moving from network plinary collaboration that systematically Universidad Cato´lica del Ecuador, spoke assembly of genome-scale data to network- generates antibodies, producing about 200 about results from frog studies (The based study of disease. He described how clones and 2 Terabytes of data per week. comparative analysis reveals independence of disease networks can aid diagnosis and Webb Miller, who received the ISCB developmental processes during early development provide a functional separation of disease Senior Scientist Accomplishment Award in frog, http://ff.im/4EvDU). del Pino’s gene families, using as an example his work for 2009, gave another keynote (Bioinfor-

PLoS Computational Biology | www.ploscompbiol.org 3 January 2010 | Volume 6 | Issue 1 | e1000640 matics Methods to Study Species Extinctions, automated motif discovery when no previ- speakers face-to-face, and enjoy the stimu- http://ff.im/4HjYa). In accordance with ous knowledge is assumed. Here, calculation lating buzz of activity. Indeed, the atten- the theme of his ISCB award, he began by of probabilities of motifs given query text, an dance at ISMB/ECCB 2009 was very close discussing his 10 steps to success in bioinfor- IUPAC motif and a random text model can to the levels expected by the organizers. matics (http://www.iscb.org/iscb-publications/ be difficult due to the large number of Organizers may also have been concerned index.php?option=com_content&view= available motifs. In the highlights track, Eric that live blogging might produce uninfor- article&id=324). Specifically, his research Alm of MIT spoke about ecological and mative content. However, this year’s cover- focuses on how genomic properties are genetic diversity (Modeling Ecological and age at the ISMB/ECCB was very sober and related to the extinction, or risk of extinction, Genetic Diversity in Bacteria, http://ff.im/ largely efficient. Misuse of officially sanc- of a species. Miller has performed sequence 4yWcn). Alm and his group, working with tioned microblogging areas, such as the analysis for a number of rare or extinct Martin Polz at MIT, collected four buckets addition of spam or rude comments, was species including the Tasmanian tiger, the of seawater off the coast of Massachusetts, another unfounded concern at ISMB/ woolly mammoth, and the Tasmanian devil. two in the fall and two in the spring. They ECCB 2009, where all comments were For extinct species, the mitochondrial ge- noticed that the organisms clustered season- pertinent, respectful, and polite. nome is up to 1,000 times more abundant ally and discovered that the free-living than genomic DNA and is therefore easier to (small-particle) organisms clustered separate- Summary sequence. Even with the limitations in ly from those attached to zooplankton. working with ancient DNA, information Many of the groups consisted of only Vibrio In last year’s report on microblogging can be discovered on phylogenetics, popula- splendidus, whose various ecological prefer- ISMB 2008 [1], the authors anticipated that tion genetics, the evolution of function, and ences were strong enough to trigger speci- new methods of using the Web and of evolutionary rates. For example, Miller ation. The samples showed the beginning of reporting the conference would make live discovered that the Tasmanian Tiger di- the sympatric speciation process. They blogging even easier (http://www.bork. verged from the Tasmanian Devil about sequenced 100 genomes from the samples embl.de/,jensen/ismb2008/keynotes.php. 40 million years ago, and that an epidemic and used their STARRIniGHTS software html). This year, the ISMB/ECCB 2009 was the probable cause of their extinction in to infer recombination breakpoints. Alm Web site contained all of the features of the 1936. discovered that most of the support for the mock-up, far more live bloggers participated zooplankton/free-living split is on an 18 kb than last year, and there was increased Overall Coverage operon responsible for zooplankton chitin coverage of talks and special sessions. We, in catabolism. turn, look forward to the new technologies Based on the statistics from the Friend- appearing on the horizon (such as Google Feed room, keynotes were the most com- Experiences Wave), and how both tools and bloggers will mented talks, likely due to the larger number make next year’s conference an even greater of attendees present (http://www.iscb.org/ Based on the organizer and attendee success. cms_addon/friendfeeds/ff_stats.php). How- experiences at ISMB/ECCB 2009 as well In summary, conference organizers ever, live bloggers attended all tracks, as at earlier conferences, we have created found that microblogging added value satellite meetings and special interest groups a set of guidelines that can be used in the for all conference attendees, and allowed (SIGs), with only a single special session creation of live blogging policies for any attendees to follow the thoughts of others missed. Five out of the nine SIGs and four life sciences conferences. These guidelines as well as to follow presentations that out of the six special sessions had greater are available in a Perspective article conflicted with others they wished to see. than 50% talk coverage. The technology published in concert with this report [2]. The usefulness of live blogging extends track, which dealt with advances in imple- Both local and remote attendees found the beyond the duration of the conference, mentations and software, was the least well live blogging helpful, informative, and a remaining accessible long after the confer- covered. Even so, three of the five most-liked useful archive, and these guidelines should ence has closed. presentations overall were from this track. help support and direct future live blog- Eamonn Maguire’s discussion of computa- ging experiences. Acknowledgments tional biology in the cloud was the third Additionally, initial ISCB concerns over most-liked talk overall (http://ff.im/4Evkq). having an open live blogging policy were We would like to thank ISCB and the ISMB organizers for supporting and facilitating our Janet Thornton’s overview of the ELIXIR allayed after the success at this year’s coverage, and all who participated in live blogging infrastructure for biological information in conference. Organizers initially feared that the conference. We would also like to thank the Europe (http://ff.im/4yWaK) and Frank live blogging might distract from the staff of FriendFeed for their useful technology and Tanoh’s presentation on the Web service conference. However, while the distraction in particular Casey Muller and Bret Taylor for registry BioCatalogue (http://ff.im/4Evnh) of an Internet connection is one that exists their rapid response to the automatic blacklisting placed fourth and fifth, respectively. whether or not live blogging is allowed, live of the conference venue’s IP address. We also Tobias Marschall’s discussion, Efficient blogging caused no discernible upset in the thank Shirley Wu, Neil Saunders, and Frank Exact Motif Discovery, was the most-comment- audience. Further, live blogging, while a Gibson for providing feedback on the manuscript. RSD wishes to thank her PI, Kimmen Sjo¨lander, ed presentation in the proceedings track useful tool for participating in topics of for supporting her attendance at ISMB/ECCB (http://ff.im/4Evno). Marschall described interest irrespective of physical presence, is 2009. ALL wishes to thank Tom Kirkwood and how measuring and locating over-represen- no substitute for being at a conference Anil Wipat for supporting her attendance at tation are the main issues in unsupervised where one can network, question the ISMB/ECCB 2009.

References 1. Saunders N, Beltrao P, Jensen L, Jurczak D, Krause 2. Lister AL, Datta RS, Hofmann O, Krause R, put Biol 6(1): e1000563. doi:10.1371/journal. R, et al. (2009) Microblogging the ISMB: A new Kuhn M, et al. (2010) Live coverage of scientific pcbi.1000563. approach to conference reporting. PLoS Comput conferences using web technologies. PLoS Com- Biol 5: e1000263. doi:10.1371/journal.pcbi.1000263.

PLoS Computational Biology | www.ploscompbiol.org 4 January 2010 | Volume 6 | Issue 1 | e1000640

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Lister, AL; Datta, RS; Hofmann, O; Krause, R; Kuhn, M; Roth, B; Schneider, R

Title: Live Coverage of Intelligent Systems for Molecular Biology/European Conference on Computational Biology (ISMB/ECCB) 2009

Date: 2010-01-01

Citation: Lister, A. L., Datta, R. S., Hofmann, O., Krause, R., Kuhn, M., Roth, B. & Schneider, R. (2010). Live Coverage of Intelligent Systems for Molecular Biology/European Conference on Computational Biology (ISMB/ECCB) 2009. PLOS COMPUTATIONAL BIOLOGY, 6, (1), PUBLIC LIBRARY SCIENCE. https://doi.org/10.1371/journal.pcbi.1000640.

Persistent Link: http://hdl.handle.net/11343/259619

File Description: Published version License: CC BY