Bohemian Bashfully. Baroque birthdays and bingo in Prague.

Every summer we celebrate our birthday at the ISMB (Intelligent Systems for Molecular Biology) conference, and this was a particularly memorable and important one – being the 5th birthday of GigaScience, and 25th edition of the ISMB.

Having ISMB in Prague this year, the historic capital of old Bohemia was particularly appreciated by the attendees, both for its beautiful baroque streets, and its reputation as a party capital with the highest per-capita beer consumption. The five day meeting lived up to all these stereotypes and more, attendance was up a lot compared to last years conference in DisneyWorld (see our write-up), and our party was a particularly lively and difficult one to top in the future. The incredibly cheap beer meant our 5th birthday budget went 5 times further than previous years, although the thirsty computational biologists did a good job drinking the venue dry of dark beer. Following last years Game of Omes theme we brought new “Data Is Coming“ t-shirts featuring our GigaPanda mascot evolving to a new “direpanda” sigil, alongside a special 5th anniversary issue that was printed for the event. We also published a 5th birthday editorial in the journal, looking back and looking forward, that you can read here.

Goodbye SIGs, hello COSIs. Big changes to the conference.

Being the 25th ISMB brought big changes for their milestone year, with a major rearrangement and restructuring of the format so that the separate SIG (special interest group) sessions that were before the main meeting, where now integrated into the main conference, and turned into COSIs (communities of special interest). This pushed the price up for people that normally only attended the SIGs, but widened the audience for the specialist communities, and also reduced the clashing between SIGs by 50%. Generally the new format seemed to be popular, although the big fallout from this may be our favorite SIG, BOSC – the Bioinformatics Open Source Conference, potentially leaving. It was announced that next year BOSC and the Galaxy Community Conference (another favourite meeting of ours that we attended a few weeks ago) will collocate as the Bioinformatics Community Conferences in Portland, shortly before the main ISMB meeting in Chicago. While this is officially only a one year experiment, depending how it goes next year the future of the meeting at ISMB is uncertain.

As sponsors of both BOSC and GCC this year we will likely have to spend time with both of our divided families next year. If this is a last hurrah for BOSC at ISMB2017, it at least went with a bang. The venue was full to bursting, and the number of new faces made up at least 30-50% of the audience. With our open science ethos we were very happy with the open science theme this year, and the keynotes were particularly great (we’ll update this with the videos when they get posted online). See the storifies of day one, two, and the many blogs (e.g. from Brad Chapman and Biologeek).

On top of open science, their tracks covered another of our big interests: workflows and reproducible research. This was the third birthday of Common Workflow Language (CWL) which came out of the 2014 hackfest, and there were a number of talks from the SevenBridges RABIX workflow system and Galaxy. As well as Stian Soiland-Reyes presenting the impressive Common Workflow Language viewer that moves the sharing of CWL workflows into the territory of myExperiment (slides here). What was impressive this year was the organic growth of open source communities behind many of these, the PIs and core team members of CWL and Galaxy could sit in the audience and let others present their efforts.

We’ve been big promoters of citizen science, published many papers promoting these projects and even crowdfunding our own Bauhinia Genome project (see Scott’s recent TEDx talk), so it was great to see a citizen science track in BOSC. As a crowdfunded project travel covered by OBF travel grants to allow Jonathon Sobel to fly from the Lausanne Hackuarium biohacker community, BeerDecoded kicked the track off talking about their community genome project to educate people about open science and genomics(video here). These were exactly the reasons we set up our own community genomics project in Hong Kong, using the Hong Kong public’s interest in their mysterious floral emblem as a hook rather than the Swiss public’s interest in beer. Like our efforts promoting reproducible research by publishing our data and protocols, they’ve been sharing their methods and data via GitHub.

Both keynotes this year covered open science, Madeline Ball talking about her work with the personal genome project and now Open Humans (video here). Inspired to get into open science through being a Wikimedian, coming out of the George Church Lab she was the manager of the original Personal Genome Project that recently published 174 haplotyped genomes with us, and with open humans now provides a one-stop-shop for consent and recruitment to donate all your medical data to medical research. Nick Loman’s talk was a love letter to ubiquitous sequencing and the Oxford Nanopore portable sequencer (see above for the video). Moving far beyond the early work he published with us presenting the first E. coli genome sequenced on the nanopore, Nick recounted us with his adventures targeting the Ebola and Zika outbreaks armed with a mobile sequencing lab. Described by Nick as being a bit like Top Gear, but without the misogyny and racism, the Zibra project saw Nick drive around Brazil in a camper van full of nanopores collecting sequencing data to determine the natural history of the outbreak. Zika being very difficult to tackle with amplicon sequencing, key to getting the data to work was Nick’s student Josh sharing the protocols, getting other to adapt and share them back, and the amount of public domain zika sequencing data has grew exponentially as a result.

The traditional BOSC panel discussion this year covered Open Data, and featured both keynote speakers, Andrew Su, and our Ed Board Member Carole Goble. Nick was very gracious, highlighting our open science efforts, and saying the crowdsourcing of the genome of the deadly 2011 E. coli outbreak that partly came out of the release of our first dataset in GigaDB (see “notes from a tweenome” for the full story) was the even that first got him into Open Science. Using our data to assemble the genome and release the results via his blog helped him become an author of the resulting “open source genomics” paper in the New England of Medicine, and ever since he has only seen positives rather then negative effects from early open data sharing on his career.

It was frustrating the BOSC Open Data track clashed with the BD2K Open Data track, but this shows that even if BOSC leaves ISMB there are plenty of other open source advocates at ISMB that can fill the gap. Overall the new COSI schedule meant there were less clashes than previous years, and it was the first time we could attend the Bio-Ontologies COSI/SIG in a number of years. In an ISMB of many birthdays, Bio-Ontologies celebrated the milestone of it being their 20th edition at ISMB, and had a looking forward/looking back panel covering where the field has gone. Memorable for Carole Goble’s interjection that ontologies are like underwear – essential but if everything is working properly then purposefully discrete – that people kept coming back to in the rest of the conference.

FAIR bingo

Like Carole’s ontology underwear comment, over the years there have been many recurring topics and memes at ISMB. In the past these have inspired bioinformatics bingo cards. Where the NHGRI/Moore’s law sequencing data graph was ubiquitous in years past, the last few years have seen the rise of docker containers, the big trend this year was FAIR bingo – where the speakers mentioned the mention of FAIR principles in nearly every talk. Having been included on the original paper was good to see FAIR (findable, accessible, interoperable, and reusable) moving beyond just a mnemonic to being implemented in practice in a number of talks. Including the Galaxy Training Network talk for Bérénice Batut at BOSC, as well from members of the GO-FAIR initiative (which is a proposal for the practical implementation of the European Open Science Cloud), Barend Mons and Marco Roos both presenting on it and showing the same FAIR video.

Drink the ELIXIR

The ELIXIR track on the last day was another #FAIRbingo-heavy session, and involved the launch of the new ELIXIR list of Core Data Resources and Deposition Databases. Presented by Jo McEntyre, the curated list gathers together the key data resources that are of fundamental importance to the wider life-science community and the long-term preservation of biological data. Being ISMB old-timers we’ve seen efforts such as BioCoreDB (attending and writing about the 2011 ISMB workshop here) working on an information specification for biological databases, but this goes beyond this to gather metrics and indicators to determine how key these resources are.

The meeting ending on a high with an amazing keynote/prize lecture from our Ed Board Member Pavel Pevzer, drawing on his incredible work moving bioinformatics education from the classroom, to online with MOOCs (massive open online courses) and now MAITs (massive adaptive interactive text). It was a great note to close on, showing that bioinformatics is getting more democratic, open and FAIR, so even if BOSC doesn’t return to ISMB in 2019, the conference will continue with the legacy of its open source ethos.

Further Reading
1. Scott C. Edmunds, Nicole A. Nogoy, Hans Zauner, Peter Li, Christopher I. Hunter, Xiao Si Zhe, Laurie Goodman; Looking back: forward looking. Gigascience 2017 gix064. doi: 10.1093/gigascience/gix064

UPDATE 3/8/17: now with added BOSC videos.