• Home
• Research
• Teaching
• Publications
• Biosketch

Research

(More in our lab homepage)

The general research objective of our lab is to develop machine learning algorithms to extract and integrate subtle and elusive information hiding in genome-wide large-scale biological data for understanding the association between genomic characteristics and phenotypes. We are particularly interested in designing theoretically principled methods in the categories of kernel methods, graph-based learning algorithms and various other statistical models for a unified analysis of the high-throughput data in a data-driven perspective: learn from known data accurate predictive models and discover from known data essential and key elements to characterize and predict phenotypes. Our current projects center around the following topics,

• Cancer genomics: development of graph-based learning algorithms for mining biomarkers of disease-related phenotypes from microarray gene expressions, SNPs and DNA copy number variations.
• Comparative genome analysis: application of network-based learning algorithms for a comparative genome study of multiple species.
• Protein analysis: development of kernel-based learning algorithms to infer the correlation between protein-protein interactions, protein structures and functions.