
243f5afc5295f7e5f42e5402187500dc.ppt
- Количество слайдов: 28
bio. Widgets: Visualization Components for Bioinformatics 11/11/97 DOE-HGP: bio. Widgets
Widgeteers G. Christian Overton, Ph. D. Jonathan Crabtree, Steve Fischer, Mark Gibson Center for Bioinformatics University of Pennsylvania 11/11/97 DOE-HGP: bio. Widgets
Background • No shortage of tools for genomics. – Contig assemblers. – Gene finders. – Sequence/motif search programs. – Databases galore. • Less emphasis on integrated systems and solutions, at least in the public domain. 11/11/97 DOE-HGP: bio. Widgets 3
Why? • Contributing factors; systems – Must be tailored to location/situation/resources. – Take more effort to build. – Require robustness. 11/11/97 DOE-HGP: bio. Widgets 4
A Solution • Provide components to build systems. • Initial focus on graphical user interfaces: – Data will soon be piling up. – Tools available for analysis. – The human expert is the bottleneck. – Enable visual “data-mining. ” 11/11/97 DOE-HGP: bio. Widgets 5
bio. Widget Principles • Java™ for Web distribution. • Focus on visualization. • Goals of component-based approach: – Enable software reuse. – Higher quality end result. – Promote development of standards; – Framework for communication. 11/11/97 DOE-HGP: bio. Widgets 6
What will be provided? • “Shrink-wrapped” applications. – Target: end-users. • Application toolkits. – Customize and build applications. – Plug-and-play model for component software. – Library routines: weight matrix, translation, etc. – Target: system builders. 11/11/97 DOE-HGP: bio. Widgets 7
History (abridged) • Postscript – “Doing Sequence Analysis with your Printer” – Searls, ‘ 93 CABIOS • Tcl/Tk and Perl/Tk bio. Widgets. • bio. Widgets meet Java™. 11/11/97 DOE-HGP: bio. Widgets 8
Current bio. Widgets • Used in-house: – Sequence (DNA and polypeptide. ) – Map (STS-content, radiation hybrid, etc. ) • Prototype: – Chromosome. – Sequence alignment. – Multiple map. 11/11/97 DOE-HGP: bio. Widgets 9
Sample Applications • Examples illustrate: – What the widgets do? – What comes in (data input)? – What goes out (output)? • All can be found on-line at: – http: //www. pcbi. upenn. edu 11/11/97 DOE-HGP: bio. Widgets 10
What comes in: GAIA • • Automated annotation of “raw” sequence. System comprised of “sensors. ” Database engine provides persistence. Visualization: map and sequence widgets – Sharing a common data source. – Communicating selection state and position. 11/11/97 DOE-HGP: bio. Widgets 11
What goes out: TESS • Exporting data to external applications. • Limited by Java™ state-of-the-art. • Soon to be improved. – True “cut and paste” both between widgets and the outside world. – Support for printing. 11/11/97 DOE-HGP: bio. Widgets 13
11/11/97 DOE-HGP: bio. Widgets 15
BLAST Visualizations • Map: Toxoplasma gondii EST database. – (24 -hour data minimart) – EST clusters fed into BLAST. • Sequence. • Standalone prototypes. 11/11/97 DOE-HGP: bio. Widgets 17
11/11/97 DOE-HGP: bio. Widgets 19
Architecture • Smalltalk Model-View-Controller. – Model = your hard-earned data. – View = the bio. Widgets. – Controller = additional bio. Widget components. • Update propagation. – Multithreaded. – Can be multidirectional. 11/11/97 DOE-HGP: bio. Widgets 20
Standards • OMG Life Sciences SIG (CORBAMed). – http: //lsr. ebi. ac. uk • Compatibility with existing standards. – Object-oriented approach. – CORBA/RMI. – Flat-file: GCG, FASTA, etc. • Must accommodate coexisting standards. 11/11/97 DOE-HGP: bio. Widgets 21
Java. Beans™ • Sun’s component architecture. • Framework for generic application components. • Allows one to create and customize applications in an “application builder”. • bio. Widget architecture is an extension. 11/11/97 DOE-HGP: bio. Widgets 22
Release Schedule • December 10: Java. Bean component release. – Tutorial at NMHCC Bioinformatics conference. – “In time for the holidays. ” • Why a Java. Bean release? – Solves problem of customization without source code. – Strikes balance between delivering complete applications and delivering everything. 11/11/97 DOE-HGP: bio. Widgets 25
Related Work • bio. Widget Consortium. • goodman. jax. org/projects/biowidgets/consortium/ • • EBI Hyperbolic Viewer. JADE. GDB Map. Viewer. Neo. Morphic Genome Browser. 11/11/97 DOE-HGP: bio. Widgets 26
Future Directions • New widgets and data models. – Database schemas. – Gene expression arrays. – Pathways (metabolic, signaling, etc. ) • One-to-one correspondence. • Need to speak the same language. 11/11/97 DOE-HGP: bio. Widgets 27
Questions? 11/11/97 DOE-HGP: bio. Widgets 28
243f5afc5295f7e5f42e5402187500dc.ppt