Butler: Analysing Genomes in the Cloud
Ask Jeeves may be a thing of the past as there’s a new butler in town. Specifically, one to serve your genomic data analysis needs.
This week scientists from the European Molecular Biology Laboratory (EMBL) have unveiled a new computational tool to facilitate large-scale genomic analysis on the cloud. This, they say, will overcome stability and efficiency drawbacks of current approaches.
In large-scale analysis projects heterogenous data are susceptible to biological artefacts or contamination which can cause anomalies in the software, hindering or failing analyses entirely. Not only does the software itself limit analyses but running cloud computing on this scale means it’s prone to system crashes. Delays in recognising these anomalies impacts the cost and duration of many projects.
Butler differs in its ability to discover and resolve these anomalies on two levels. Firstly, it manages virtual devices by continually collecting system health data from all it’s components. It’s then able to “self-heal” by taking automated steps to restart failed elements. Additionally, the monitoring system alerts users, via email or the messaging platform Slack, when running normal conditions are breached – for example when automatic action is ineffectual. By providing researchers with the tools for troubleshooting the software it reduces the technical competency required by the researchers to run genomic analyses.
It’s these features that allow this workflow to be highly efficient when used by multiple groups on different clouds. As Butler was trialled in the Pan-Cancer Analysis of the Whole Genome Study (PCAWG) it was essential that the multi-cloud functionality was flexible to handle the variety of regulatory requirements of the international collaborators. As Butler is able to run on most academic and commercial cloud platforms it holds genuine potential for accelerating global big data analyses.
For those interested in data driven drug discovery/development, join us in Basel later this year at D4 Europe. Find out more about this event and register before the super early bird ends on Monday 24th February www.d4-europe.com