geWorkbench

1
geWorkbench

• John Watkinson
• Columbia University

2
geWorkbench

• The bioinformatics platform of the National
  Center for the Multi-scale Analysis of Genomic
  and Cellular Networks (MAGNet).
• Also, part of the NCIs cancer Biomedical
  Informatics Grid (caBIG) initiative. The project
  was formerly called caWorkbench.

3
geWorkbench (cont.)

• A desktop application for integrative genomics.
• Runs on Windows, Linux and Macintosh.
• Includes a variety of informatics tools, but
  specializes in microarray analysis.
• Open-source and free for non-commercial use.
• Includes an API for plugin development.

4
(No Transcript)
5
geWorkbench (cont.)
6
Integrative Genomics

• Increasingly, researchers need to combine several
  data sources (microarray assays, DNA/RNA/protein
  sequences, protein structure, gene ontology,
  clinical data, etc.)
• geWorkbench attempts to move past simple
  microarray analysis to include integrative
  methods.
• Plugin framework allows geWorkbench to interact
  with other major software packages, including
  BioConductor, GenePattern and Cytoscape.

7
Data Support

• Microarray assays (one-color and two-color, as
  well as caARRAY assays).
• Sequence files.
• BLAST queries.
• Gene-Gene interaction networks (Interactomes).
• Gene Ontology Terms.
• caBIO pathways and annotations.
• Protein structure files (PDB).

8
Components

• geWorkbench has a plugin interface for the
  development of 3rd-party components.
• Documentation and developer support is available
  from the geWorkbench team.
• All visualizations and analyses have been written
  using the API. Several groups at Columbia are
  developing for the platform.

9
Microarray Analysis

• Summarization of raw chip data (via
  BioConductor).
• Normalization and Filtering.
• Differential expression analysis.
• Clustering (Hierarchical and Self-Organizing
  Maps).
• Classification (SVM and SMLR).
• Many visualization tools.

10
Hierarchical Clustering
11
Scatter Plot Visualization
12
caBIO Pathway Viewer
13
Sequence Analysis

• BLAST and HMM search interface.
• Pattern discovery.
• Synteny analysis.
• Promoter region analysis.
• A variety of sequence viewers.

14
Pattern Discovery Viewer
15
Promoter Viewer
16
GO Term Enrichment

• Traditional t-tests on microarray data determine
  differentially expressed genes between two
  different phenotypes.
• Gene Ontology (GO) term enrichment can determine
  which functional or structural categories show
  significant differentiation.
• Supported in geWorkbenchs GO Panel component.
• A similar technique can be applied to other gene
  sets, such as KEGG pathways.

17
GO Terms (cont.)
18
Reverse Engineering

• Microarray data can be used to infer biological
  pathways.
• geWorkbenchs Reverse Engineering component uses
  the ARACNE algorithm to build gene-gene
  interaction networks.
• These can be compared and combined with an online
  database of interactions, curated by Columbia.

19
Reverse Engineering (cont.)
20
Reverse Engineering (cont.)
21
Matrix REDUCE

• Given microarray data and upstream sequences for
  genes, transcription factor binding sites can be
  inferred.
• The Matrix REDUCE component in geWorkbench
  provides this analysis and tools to visualize the
  results.

22
For More Information

• http//www.geworkbench.org
• Mailing List
• geworkbench-users_at_gforge.nci.nih.gov
• John Watkinson watkin_at_c2b2.columbia.edu

23
Acknowledgements

• ARACNE algorithm by Califano et al.
• Matrix REDUCE algorithm by Bussemaker, et al.
• geWorkbench team Aris Floratos, Eileen Daly,
  Kenneth Smith, Kiran Keshav, Xiaoqing Zhang,
  Manjunath Kustagi, Matthew Hall, Bernd Jagla,
  Mary VanGinhoven, John Watkinson.