Speed Group
Microarray Page
Index to our site
Title: Discrimination and Clustering with Microarray Gene Expression Data.
Authors: Terry Speed, Jane Fridlyand, Yee Hwa Yang, SandrineDudoit.
Place: ENAR Conference
Date: March 27, 2001
Abstract:
Clustering methods are very widely used in the analysis
of microarray gene expression data, at times when it is more appropriate
to use some form of discriminant analysis or an even simpler comparison
of means. However, clustering genes followed by averaging expression levels
within clusters can reduce noise and permit more sensitive comparisons.
Moroever, clustering of samples (e.g. from tumor tissue) can reveal previously
unrecognized subclasses. Thus there will remain a valuable role for clustering
in microarray data analysis, although perhaps not as broad as currently
perceived. One question that then arises is how many clusters do we have,
and how reliably can we allocate units to clusters. In this talk I outline
some of the uses of discrimination and clustering, and touch on some novel
approaches we are exploring to the questions mentioned.
Slides: Download power point file [ppt file][jpeg files]
Last modified: Sat Apr 7 12:54:33
PDT 2001
zarray@stat.berkeley.edu
