By Shoaib Sufi, Community Leader
The Collaborations Workshop 2015 (CW15) and Hackday lasted three short - but highly charged - days, and attracted over 85 people to work on interdisciplinary research. As the rich set of resources created at the event are written up, we will be releasing these short posts to share the outcomes and tools with the community.
The CW15 demos covered a vast array of subjects. From systems for creating data management plans (DMPonline), to new ways of packaging software, data, papers and other associated resources to make Research Objects the new index of your research work. Other demos showed off tools to help visualise data sets using Python (DAWN Science), workflow-oriented systems that make it easier to integrate the vast array of web-based datasets using a visual programming paradigm (such as Apache Taverna, and systems for cataloguing data driven experiments (SEEK).
Bioinformatics tools, services and know how had a huge showing, with strong representation from the Wurmlab. Important approaches to software development for researchers were covered such as the user centric design approach used to develop sequence server. The need to identify and fill in the gaps in software provision in Bioinformatics was highlighted with an exemplar of the need to validate gene predictions and how this lead to the Gene validator software system. Techniques such as crowd sourcing applied to improving the quality of research data offered a novel example of how techniques from other areas could be used to improve the quality of research tools, such as the gene prediction system, Afra).
The explosion of genomics data is one of the key examples of Big Data. Tools to handle the ever burgeoning size of data in the Life Science are always well received. Such a tool was the Bionode system. Bionode is inspired by the Unix pipe’s philosophy and offered a stream based way of processing large data sets, this really decreased memory requirements for handling large data files and is an exciting proposition for dealing with larger data on a routine basis with standard hardware.
There was much talk of container-based approaches to aid reproducibility. Containers are viewed favourably due to their lower space, memory and time requirements compared the alternative method of using virtual machines to package tools, experiments or systems. Of specific interest at the Collaborations Workshop was Docker. One of the systems developed at Wurmlab made it even easier to use Docker to aid reproducible research, solving the problem of differing bioninformatics requiring different libraries but needing to operate in the same computer environment they demoed the OSwitch system which made this inspirational goal a reality.
Research is not a purely technical field, but has strong social aspects; a socio-technical environment. In this regard Derek Groen had run a successful Paper Hackathon where coders, researchers and data mixed to produce a real research paper as an output. Derek presented his thoughts on the process and lessons learned, and he even provided a set of questions for those thinking of organising a paper hackathon of their own.
For links to the slides and more information please take a look at the CW15 Demos page.
If there are systems you would like to demo or hear more about let us know and we will try to factor that in next year at CW16.