|
Activity Number:
|
143
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308498 |
|
Title:
|
Identifying Differentially Expressed Genes for Time-Course Microarray Data Through Functional Data Analysis
|
|
Author(s):
|
Kun Chen*+ and Jane-Ling Wang
|
|
Companies:
|
University of California, Davis and University of California, Davis
|
|
Address:
|
Department of Statistics, Davis, CA, 95616,
|
|
Keywords:
|
Time-course Gene Expression ; Differentially Expressed Genes ; Functional Data Analysis ; Functional Principal Component ; hybrid EM
|
|
Abstract:
|
In order to identify genes that show differential expression under two conditions, we develop and employ the Functional Principal Component (FPC) model to depict the dynamics of the gene trajectories. In our model, the gene expression trajectories are captured by less basis functions than the other methods, and these basis functions are estimated from the data reflecting major mode of variation in the data. The dependency structure of the gene expression over time is also considered without any parametric assumptions and estimated from all genes such that the information across the other genes can be borrowed. The parameter estimation is carried out by hybrid EM algorithm for an efficient computation. We apply our method to real and simulated data set and compare it to two-way mixed ANOVA method. With little model assumptions FPC analysis shows better performances.
|
