|
Activity Number:
|
391
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #300530 |
|
Title:
|
A Gene Selection Method for GeneChip Array Data with Small Sample Sizes
|
|
Author(s):
|
Zhongxue Chen*+
|
|
Companies:
|
The Children's Hospital of Philadelphia
|
|
Address:
|
Biostatistics and Data Management Core, philadelphia, PA, 19104,
|
|
Keywords:
|
Multiple comparison ; False discovery rate ; Microarray ; gene selection
|
|
Abstract:
|
Multiple comparison issue is a challenge in gene selection of microarray data. Current methods that control FDR are not applicable when sample sizes are very small due to the poor estimations for p-values and the distribution of true null. We propose a model-based information sharing method, which takes the advantage of information shared among genes and has better test statistics. After appropriate data transformation, we use a normal distribution to model the mean differences of true nulls. The parameter in the model is estimated by all data. The p-values are calculated from the model. Instead of controlling FDR,we select genes for a given cutoff p-value and then estimate the false discover rate. We compare our new method with others by simulation and real microarray data. The comparison results show that this new method is more reliable.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2008 program |
