|
Activity Number:
|
215
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 12:00 PM to 1:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section*
|
| Abstract - #301027 |
|
Title:
|
Sample Size Requirements for Comparing Two Populations with DNA Microarray Analyses of Gene Expression: A Monte Carlo Method
|
|
Author(s):
|
Craig Wood*+ and Joyce Burcham
|
|
Affiliation(s):
|
Geisinger Health System and Geisinger Medical Center
|
|
Address:
|
100 N. Academy Avenue, Danville, Pennsylvania, 17822, USA
|
|
Keywords:
|
sample size ; microarray ; Mote Carlo ; genetics
|
|
Abstract:
|
Due to lower costs in the technology, microarray analyses of gene expression can be done more frequently and at smaller institutions. As a result, statistical considerations for the design of these experiments are increasingly significant. Particularly, a simple sample size determination specifically geared toward microarray analysis of gene expression is needed for experimental design and to measure feasibility of accepting or rejecting a research hypothesis. This experiment used Monte Carlo simulations consisting of randomly generated microarrays to measure statistical power for comparing two populations. Simulations consisted of a large range of samples with properties inherited from two different populations and had a wide range of between group and within group variation. Diagnostic accuracy for each simulation was measured by comparing actual classification to the group assignment based on variance factors of the sample. Sample size nomograms are provided.
|
