GenExplore: Interactive Exploration of Gene Interactions  

Description

Microarray data provides a powerful basis for analysis of gene expression. Data mining methods such as clustering have been widely applied to microarray data to link genes that show similar expression patterns. However, this approach usually fails to unveil multiple interactions by the same gene. In this project we aim to apply statistical models (Graphical Gaussian Models, Graphical Loglinear Models) and data mining techniques (Hierarchical Clustering, Association Rules) to discover gene interactions (pairwise and multi-way levels). We aim to construct a prototype system that permits rapid interactive exploration of gene relationships; results can be validated by experts or domain information, or suggest new experiments. 

Research Tasks

• Develop a framework for interactive, scalable and exploratory gene interaction analysis.
• Investigate various interaction analysis techniques to discover gene interactions and derive genetic networks. The models include:
  • graphical gaussian modeling to discover undirected pairwise linear gene interactions for subsets with a relatively large number of genes.
  • graphical decomposition techniques to get components with a relatively small number of genes.
  • graphical loglinear modeling  to discover multi-way gene interactions within each component.
  • causal modeling to discover directed pairwise interactions.
• Investigate various visualization techniques for interactive gene interactions.
• Apply system to study prostate cancer and neuroblastoma.

     Faculty member

• Xintao Wu                 
• Kalpathi Subramanian

      Graduate Student

• Yong Ye (Ph.D. student)

• Liying Zhang (Postdoc Researcher, Memorial Sloan Kettering Cancer Center)

Papers

• X.Wu, Y. Ye. "Exploring Gene Causal Interactions Using an Enhanced Constraint-based Method", Pattern Recognition 39:2439-2449, 2006.
• Y.Ye, X.Wu "Efficient Causal Interaction Learning with Applications in Microarray", ISMIS05.
• Y.Ye, X. Wu, K. Subramanian and L. Zhang. "GenExplore: Interactive Exploration of Gene Interactions from Microarray Data", ICDE 2004 (demo) .
• X. Wu, Y. Ye and L. Zhang. "Graphical Modeling Based Gene Interaction Analysis for Microarray Data", SIGKDD Explorations, 5(2) 91-100, Dec 2003.
• X. Wu, D. Barbara, Y. Ye. "Screening and Interpreting Multi-item Associations Based on Log-linear Modeling",  SIGKDD2003.
• X. Wu, D. Barbara, L. Zhang and Y. Ye. "Gene Interaction Analysis using All k-way Interaction Loglinear: A Case Study on Yeast Data", ICML03 workshop: Machine Learning in Bioinformatics.
• X. Wu, Y. Ye, K. Subramanian and  L. Zhang. "Interactive Gene Interaction Analysis Using Graphical Gaussian Model". BIOKDD03.

GenExplore 0.98 beta version is available.

Related Links

          Some research projects, papers and products can be found here