BOSC & ISMB 2015

July 22, 2015


The Bioinformatics Open Source Conference. At around 200 people and a single track, this event is just big enough for my taste. One day you watch a speaker giving the opening keynote and the next she sits next to you tweeting away. That part I really enjoy.

This post is meant to summarise my thoughts on BOSC and ISMB. For a more detailed account on all BOSC talks, please read up on Brad Chapman's excellent notes.

Common Workflow Language

The bane of bioinformatics has to be the constant flow of pipeline solutions being thrown on the wall to see what sticks. Right now Common Workflow Language has a lot going for it. What makes me exited is to see and talk to the people that are heading the project, including the Arvados platform.

At this point, though, it feels very "alpha" and I miss a catchy demo introduction to attract more contributors and users. The reference implementations indeed works well for the demo workflows. However, looking under the hood I would hope more effort is spent to clean up and document the code.

And speaking about pipelines there is at least one unifying theme...


The drinking game keyword for BOSC was: "Docker". The same was true for PyCon Sweden earlier this year. It's definietely a technology that is pushed heavily. I think to stay on top of where the field is moving it's high time to look into deploying and using containers. Perhaps add a Dockerfile to your GitHub repos as part of your software distribution as an exercise.

It's interesting to also hear from end users like @hollybik pointing out that researchers can learn to use command line tools etc. but the added abstraction of containers might be taking things one step too far.


The panel on diversity triggered some interesting discussion and was cool to listen to. The attendee distribution in the room was exactly what you would expect. I don't doubt things will eventually change but I am concerned that it's taking an awfully long time to do so.


After 4 days of hackathon + conference I didn't exactly have the energy or exitment to fully commit to the already overwhelming ISMB conference. However, after pouring over the schedule and setting my own expectation at a reasonable level I was able to have a rather rewarding experience.


After unfortunately missing the first part of the CADD talk by Martin Kirchner. I sent an email to ask for a spontaneous meeting and ask some stupid questions. Apparently this is a perfectly valid approach and actually a much more natrual option for an introvert like myself.

Martin was very approachable and accepting of my ignorance. He took his time explaining the theory and I think I even got some valuable information regarding how to use CADD for compounds and homozygote calls. I definetly want to keep in touch.

Overall impressions

Most of the conference talks focused on stuff like expression data, transcriptomics, statistics etc. This I find much harder to catch my attention. Some topics sounded interesting but overall wasn't for me.

What gets me exited is partly learning about new or novel uses of technology including Node.js, Docker, Bionode, StackStorm, Sambamba, and CADD. Completment this with meeting the developers and users of these tools and you have me hooked. The potential to talk, learn, and collaborate is the true value of going to these conferences in my mind. And for this purpose the hackathon format is very difficult to beat.

Lastly, who knew that the best way to find collaboration opportunities within SciLifeLab was to travel to a conference abroad :smile: