26
Click to edit Master title style European Molecular Biology Laboratory, Heidelberg, Germany QuickTime™ and a TIFF (Uncompressed) decompressor are needed to see this picture. EMBL Georgios Pavlopoulos TAC-2, 15 Nov 2007 Data integration & knowledge management group Structural and Computational Biology unit A visualization tool for high level relationship and clustering analysis in large scale networks Georgios Pavlopoulos

Data integration & knowledge management group Structural and Computational Biology unit

  • Upload
    marin

  • View
    42

  • Download
    1

Embed Size (px)

DESCRIPTION

Data integration & knowledge management group Structural and Computational Biology unit. Georgios Pavlopoulos. A visualization tool for high level relationship and clustering analysis in large scale networks. Known visualization tools. Pajek. NetDraw. HyperGraph. Ondex. MultiNet. - PowerPoint PPT Presentation

Citation preview

Page 1: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Data integration & knowledge management group

Structural and Computational Biology unit

A visualization tool for high level relationship and clustering

analysis in large scale networks

Georgios Pavlopoulos

Page 2: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Known visualization toolsPajek

Medusa

Ondex

Cytoscape

MultiNet

Otter

Plankton

Osprey

NetDraw

Negopy

SocNetV

Tulip

HyperGraph

GraphViz

Page 3: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Large scale networks

What if the network is a bit bigger with many connections?

Is there any way to visualize some clusters out of this mess?

Page 4: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Motivation – General goal

1.Interactive

2.Visualize everything in 3D

3.Combine different kinds of data under the same Network

4.Provide and Visualize some clustering algorithms

Page 5: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Motivation – General goal

A C

AB

C

CB

A

5.Keep it generic so that it can be used in any case study

6.Keep it compatible with already existing tools

8.Extract indirect connections – Find hidden information

7.Maintain it read a very simple input file format

Direct connection

Indirect connectionBetween A-C

Page 6: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Arena3D

Page 7: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

…about Arena3D

Tree based clustering algorithms:

UPGMA

NJ

HCL

Non-Tree based clustering algorithms:

MCL (not yet)

Affinity Propagation

K-Means

Page 8: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Input file example

Input file example:

node_i:layer _x node_j:layer_y weight

A:pathways B:pathways 5.61

A:pathways A:chemicals 1.2

B:chemicals A:diseases 4.3

A:diseases C:proteins 2.7

Page 9: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Overview – My part

EMBL public data

Visualization

Databases Text Mining

Page 10: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Connectivity with SRS

Evangelos PafilisWeb Servises

Page 11: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

SRS: data integration system

• > 80 databases in EMBL Heidelberg

• queries against multiple databases

• cross-linking between the records

http://srs.embl.de

Venkata Satagopam

Page 12: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Connectivity with Bioalma

EMBL public data

Visualization

Databases Text MiningEvangelos Pafilis

Page 13: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

bioalma: query

Page 14: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

bioalma: analysis creation

Page 15: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

entity recognition & co-occurrences

Page 16: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

bioalma:analysis report, cooccurrences

Page 17: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

bioalma:analysis report, cooccurrences

Page 18: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Overview

EMBL public data

Visualization

Databases Text Mining

USER

Page 19: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

DEMO

VIDEO - DEMONSTRATION

Page 20: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Snapshots

Page 21: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Snapshots

Page 22: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

What I did last year

Better graphicsMore interactiveIncrease memory and speed performanceSimpler GUIDirected graphs supportMoving layers in 3D space

Clustering Algorithms – individual layersClustering algorithms – layer combinationPREDEFINED clustering

Indirect ConnectionsIntegration with SRSIntegration with Bioalma Text Mining

Page 23: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

What is next?

Data analysisMitocheckAnne ClaudeTamahud project Bioquant project

Functionality

SBML supportEven more interactivity – make everything clickableMinimization of crossoversApply the same functionality to Medusa-2DSub-Network selection

Future planPublicationEMBLEM license – invention record form

Page 24: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Page 25: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Group Members - Acknowledgements

Page 26: Data integration & knowledge management group Structural and Computational Biology unit

Click to edit Master title style

European Molecular Biology Laboratory, Heidelberg, Germany

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

EMBL

Georgios Pavlopoulos TAC-2, 15 Nov 2007

Thank you !