|
Activity Number:
|
459
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 4:50 PM
|
|
Sponsor:
|
IMS
|
| Abstract - #309703 |
|
Title:
|
Accounting for Selection Bias in the Estimation of Fold Change in Microarray Experiments
|
|
Author(s):
|
George Wright*+
|
|
Companies:
|
National Institutes of Health
|
|
Address:
|
NCI, 6130 Executive Blvd Rm 8144, Rockville, MD, 20852,
|
|
Keywords:
|
microarray ; shrinkage ; empirical bayes
|
|
Abstract:
|
In order to account for the multiple comparisons in microarray experiments it is necessary to choose extremely strict cutoffs for determining statistical significance. As a result those genes selected will tend to include those which have the greatest observed fold changes, through a combination of both chance and true biological difference. Thus the observed fold differences may be much larger than would likely be observed in follow-up experiments that attempt to verify the presented results. The magnitude of this bias depends heavily on the shape tails of the distribution of differential expression. We present several different methods to adjust for this bias and evaluate their performance under a number of simulated and experimental settings.
|
