CCBM Seminar Series

Donald Geman

Professor, Applied Mathematics and Statistics

The Johns Hopkins University

Classifying Gene Expression Profiles from Pairwise mRNA Comparisons

Statistical inference from gene expression microarray data is difficult due to the small number of observations, typically tens, relative to the large number of genes, typically thousands. Consequently, standard methods in machine learning may lead to over-fitting and inflated estimates of performance in detecting disease, identifying tumors and predicting treatment responses. Moreover, the results may be difficult to interpret in biological terms.  We address these problems by a purely rank-based analysis, comparing the mRNA counts in selected pairs, and demonstrate how this can lead to highly accurate and transparent decisions from small samples in standard classification tasks.  This is joint work with Christian d'Avignon and Dr. Raimond Winslow.

For information on disability access contact Anne Albinak at 410-516-5310 or aalbinak@bme.jhu.edu

Tuesday, March 23, 2004
4:00-5:00pm

Room 110,
Clark Hall

and

videocast to

Talbot Library,

709 Traylor Building