Activity Number:
|
394
|
Type:
|
Invited
|
Date/Time:
|
Thursday, August 15, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Quality & Productivity*
|
Abstract - #300355 |
Title:
|
Model Based Clustering and Outlier Detection in Microarray Data
|
Author(s):
|
Johanna Hardin*+ and David Rocke
|
Affiliation(s):
|
Fred Hutchinson Cancer Research Center and University of California, Davis
|
Address:
|
1100 Fairview Ave N, PO Box 19024, Seattle, Washington, 98109-1024,
|
Keywords:
|
robust ; model-based clustering ; outlier detection
|
Abstract:
|
With microarray technologies becoming increasingly accessible to researchers, more and more complex questions are being asked about the structure of the data. The tools which identify genes that distinguish predefined groups are being refined, but we are still in need of good tools that can identify groups of genes or samples without supervision (or with only moderate supervision.) Though hierarchical techniques are unsupervised, they impose a structure on the data that may not be valid. We present an algorithm for applying a model-based clustering mechanism to microarray data that encompasses an outlier detection method. We propose unsupervised gene clusters and only moderately supervised sample clusters.
|