Chipster and collaboration with other bioinformatics platforms Aleksi Kallio CSC – IT Center for Science chipster@csc. fi
Chipster introduction
Chipster in a nutshell Ø Free, open source software for analyzing high-throughput data such as NGS Ø Available as a ready-to-run VM with a large collection of analysis tools and reference data • Use directly or via Chipster GUI Ø Chipster GUI enables users to • Visualize data efficiently • Share analysis sessions • Document what they have done • Save and share automatic workflows
Analysis tools for different kinds of data Ø 140 NGS tools for Ø 140 microarray tools for • RNA-seq • gene expression • mi. RNA-seq • mi. RNA expression • exome/genome-seq • protein expression • Ch. IP-seq • a. CGH • FAIRE/DNase-seq • SNP • Me. DIP-seq • integration of different data • CNA-seq • Metagenomics (16 S r. RNA) Ø 60 tools for sequence analysis • BLAST, EMBOSS, MAFFT • Phylip
Technical features Ø Client-server user interface, loosely coupled distributed backend • Can be spread over different clouds, elasticity Ø Data on client or server side • No duplication of data Ø Workflows – reusing and sharing your analysis pipeline • You can save your analysis steps as a reusable automatic ”macro” Ø Web based interface for system administration and tool development • Tool scripts can be R, Python or Java Ø Integrated user support functionality • Easy to see what the user has done
Chipster admin GUI
Chipster compared to Galaxy Ø There is no obvious way to compare two complex systems… • Windows vs. Linux, vi vs. Emacs, Python vs. Java… Ø Many technical differences, but maybe the core difference is in how tools, workflows etc. are presented to user • Chipster’s approach is more integrated: focus on usability, consistent biologist friendly terminology, single complete virtual machine distribution, automated updates for the whole system • Galaxy’s approach is more modular: focus on tool distribution, tool developer community, workflow driven work, several customised versions available • YMMV… Ø Typical feedback we hear: in Chipster people like the GUI and in particular being able to visualise the session, in Galaxy people like the breadth of available tools and integrations
Collaboration opportunities
Tool evaluation and selection Ø Selecting best tools takes effort Ø Finding and testing example datasets takes effort Ø Wiki for shared best practices? Ø Should include basic justification for selection e. g. benchmarks, references to review articles… Ø Lightweight alternative to full comparison or review article
Cloud integration Ø Combined efforts to bring different bioinformatics platforms to major generic and scientific clouds Ø Not only about software infrastructure parts (easy), but tools and databases (hard) and keeping them up-to-date (really hard) Ø Tools to achieve elasticity not only within platforms, but across platforms Ø Running several platforms and scaling resources across the platform according to changing workloads
Chipster in EGI Fed. Cloud Ø Chipster VM available in Fed. Cloud Applications Database Ø Chipster Virtual Organization
Tool platform Ø Why every platform needs to integrate the same tools and databases, but in a different way? Ø There are many highly sophisticated solutions out there, but typically with low coverage of the day-to-day tools Ø Virtual machine images and containers (Docker) are practical tools for software packaging Ø Supporting technologies Ø Bio-Linux, Debian Med Ø Bio. Img. org Ø Cern. VM-FS
Tool platform idea Ø Factory that generates VMI’s or containers 24/7 • Always tested • Latest software versions Ø High-quality and widely used VMI’s/containers • One good software bundle is better than dozens of poorly baked ones Ø Vision: you can just assume that software and databases are there. “Spotify” of bioinformatics software. Ø For cloud: automatically updated VM with hooks for update events Ø Could there be Ne. IC project around this? 16
Finally
Thanks to users and contibutors!
More info Ø chipster@csc. fi Ø http: //chipster. github. io/chipster/
Скачать презентацию Chipster and collaboration with other bioinformatics platforms Aleksi
507d57040eb5cd9d6c9917a1b9e8d021.ppt