Generic Tools and Methods for Data Handling

1
Generic Tools and Methodsfor Data Handling

• Generic vs. Specific

2
Generic

• Models for -omics experiments
• Generate database scheme from model
• Generate XML schema from model
• Generate access code from model
• Special Interest Group Generic Tools and Methods
  Data Handling
• Using ontologies
• Investment for the future

3
Specific

• Micro Array Gene Expression (MAGE) data
• MAGE-stk toolkit
• MAGE-ML import/export
• MAGE-OM api access
• Quick and dirty

4
SIG Generic Tools and Methods Data Handling
5
Status generic

• Get the best from 3 worlds docDBcode
• Unified tools not yet found
• Evaluate part-to-part connections

Databases
X / R / OO
Codegeneration
CastorJDOHibernate
Mt, XTables
Codegeneration
DocumentsXML
Programming
Data / Doccentric
CastorJDOHibernate
6
Framework so far
VL OOprograms
code
converter
MAGE-VLfile
MAGE-MLfile
models
UML
DDL
OODBMSMatisse)
Mt.exe
XMI
ODL
XQueryBridge ?
Xqueryad hoc queries
XML Schema
GraphViz
Graphicaloutput
Tabularoutput
XformsXGUI
7
Visualization using Xquery,dotty and Graphviz
8
MatisseJ2EESVG data viewer

• Objects are Nodes
• Relations are Edges
• Looks like graph visualization with few
  algorithms and parameters
• Too many neighbors pruning
• Hubs and authorities clusters networks
• Multiple views IsA, HasA, etc.

9
Example select from Person
10
Summary, The End

• Nice ideas
• Model independent
• Re-use of tools in VLE
• Progress (too ?) slow on my own
• Crappy tools
• More resources
• Or skip parts

11
Normalization project VLE

• (too much?) Focus on generic
• Build VLE like -omics framework.
• Using Matisse OO-DBMS
• With web-access data viewer
• Java and R access
• (specific) MAGE-ML I/O
• Web-services

12
Status Matisse OO-DBMS

• MAGE-OM model to Matisse OO DBMS
• Lose some information (packages export MAGE-ML)
• Matisse is cripple-ware
• Almost no support
• MAGE-ML data converter for import
• But Matisse gives no error message

13
Status web-access data viewer

• Access with J2EEHTML in tables
• Data is too complex for tables
• J2EEForm based editor for mage-ml
• Have DTD but need XSD
• J2EESVG data viewer
• Nice, wait for 2nd presentation
• Need tighter integration
• Tools (dot) are crashy

14
Status Java and R access

• R coupling of data files via J2EEXML
• Too slow
• Java code generation
• Matisse works well
• Reflection also (everything is just a generic
  object)

15
Status MAGE-ML I/O

• Import
• Matisse gives no error message
• Export
• Lose some information (packages export MAGE-ML)

16
Status Web-services and Ontologies

• Kept in mind but no work done
• Connection with Taverna should be possible
• What the heck are ontologies?