Colloquium

Starts: October 30, 2009 at 3:00 PM
Location: Jones Hall 301
Contact: Tanujit Dey

Summary

Speaker: Yuping Wu (Cleveland State University)

Full Description

Title: Statistical Advances and Challenges for Analyzing Proteomics Data using Mass Spectrometry

Abstract: The use of mass spectrometry (MS) as a means of analyzing the proteome for biomarker discovery and cancer diagnostics has been evaluated extensively in recent years due to the new advances in MS technology.  As more MS data becomes available to researchers, significant challenges have arisen with regard to analyzing the datasets.  A key requirement for its success is the ability to identify proteins whose expression levels are different between different experimental groups (e.g. healthy vs. diseased) in complex mixtures.  In this talk, I will first introduce procedures for pre-processing of mass spectra, including steps of denoising, baseline correction, normalization, peak extraction, peak alignment and some among others.  Then I will propose a novel locally adaptive discriminant analysis technique for cancer classification and biomarker discovery based upon location/intensity information associated with peaks.  The methods will be evaluated via applications to simulated data using the virtual mass spectrometer along with various forms of existing proteomics data, available in the literature and in the public domain.