Friday, 22 July 2011

2011 Codefest, BOSC and ISMB

Vienna just hosted the Codefest of the Open Bioinformatics Foundation, followed by the Bioinformatics Open Source Conference (BOSC) and the main conference Intelligent Systems in Molecular Biology (ISMB). Debian Med had a talk+poster at BOSC and somehow nobody took that down for the main conference which (for some this may be the proof for the existence of some higher might) let it hang next to the only poster by Microsoft among some 1000 or so. Hillarious. It should be noted that the Microsoft representative, Dr. Simon Mercer, seems to be a nice guy and he did not take it down either.

As a community it was nice to meet many contributors to Debian again, bringing up sweet memories also from January's sprint and Tim to volunteer organising the next one. The Codefest brought preliminary packages for GMOD's chado and jbrowse in collaboration with Scott Cain, who later at BOSC finally met with Olivier, our maintainer of gbrowse. One was also reminded of a series of nice efforts to bring microscopy and its analysis of cellular images closer to Java. And besides ImageJ this also means the need for Fiji. To bring our medical and biological users closer together, those packages have some extra importance to us. After some exchance with respective upstreams, things are getting increasingly ready for our (re)distribution. Michael from the Ensembl helpdesk had a review of the Ensembl packaging and he liked it a lot, which is good. Well, this is not too surprising, since it is maintained by Eagle Genomics - professionals in Ensembl and Cloud Computing ... and sponsors of BOSC. He also encouraged us to submit patches to remove/reduce dependencies on the long outdated version 1.2.3 of BioPerl, so it could leave the experimental section. Jim Procter sees some light towards a Debian package of Jalview. BOSC then brought a series of additional promising contacts and so did the ISMB itself. There is a strong interest from many sides to get getData into a pristine shape. With Peter Rice we have now agreed to get instructions for his EMBOSS tools, maintained by Charles, into it, starting with the classical protein sequence, motif and interaction databases. One is tempted to bring also a text based human genome in from Ensembl, but we'll see how far we get.

A major dominance throughout the Codefest and BOSC in a series of presentations had the Cloud Bio-Linux initiative. Everybody seems to like it, also in industry. It will be described in more detail in a separate post. It deserves it. When Debian approaches the best possible bioinformatics environment (what is that? and for whom?) in a bottom-up approach, Cloud Bio-Linux is top-down: bring something up that works and use Debian technologies whenever this his helpful. They will ask for money, from the government and from industry, earmarked for their bioinformatics ambitions, and from those positive vibes on the conference one tends to think the'll get some. When remembering e.g. the Dunk Tank disaster, anything like that seems rather unlikely to happen from the Debian side. But that is fine. Cloud Bio-Linux is contributing to Debian, and their contributors are feeling to be a part of Debian Med, too.

The community is possibly the most important aspect of it all. The expertise on biomedical computing collected in a single room at the Codefest and BOSC is enormous. And while there are many many many many Macs on many laps, about everyone one talks to uses Ubuntu on their
desktop/servers back home. The Debian Med's repository gives them all an opportunity to bring their expertise up into a collaborative environment and form something larger from there. Last and this year this is the cloud and what you could do with it. Next year brings complete tangible workflows, we are sure.