NYU Microarray Database (NYUMAD)

1
NYU Microarray Database (NYUMAD)

• Marc Rejali

2
Overview

• Supports collaborative research
• Large data sets
• Extensible Analysis functionality
• Efficient and intuitive interface
• Flexible design
• Use of standards (MAGE-ML)

3
Architecture/Design
Front Tier
XML ( MAGE-ML)
Local C or C modules on users computer
HTTP
Middle Tier
C, C, Java, other languages
Back Tier
4
Data Sharing

• Collaborative Data sharing
• Controlled visibility/access
• group-based read/write access
• public/private data
• Community access to public data
• HTTP communication
• XML and MAGE-ML standard
• Easy data import/export
• Can create customized data sets
• Can filter and transform export data

5
Array Definition Screen
6
Experiment/Hybridization Data Screen
7
Data Analysis 1

• Clustering analysis
• Data filtering, normalization, transformation
• New shrinkage similarity measure
• Gene and experiment cluster generation
• Graphical display of clustering
• Expression profile analysis
• Find expression profiles correlated/anti-correlate
  d to a given genes expression profile
• Compare to a synthetic profile
• Data filtering, normalization and transformation
  available

8
Data Analysis 2

• Use of fast plug-in C/C analysis modules
• Sophisticated new technique for normalization and
  differential expression

9
Clustering Screen
10
Data Content

• Experiments
• calculated and typed experimental factors
• templates for factor input
• Hybridization results
• batch and regular import
• Array Design
• array feature annotations
• User arrays
• Protocols

11
Software Engineering

• Three-tier architecture ensures performance,
  scalability and flexibility
• Client and Server caching ensure speed
• Re-use demonstrated with NYUSIM