|
Activity Number:
|
543
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
SSC
|
| Abstract - #302972 |
|
Title:
|
Survival Relevant Gene Selection in Microarray Data Analysis with Gene Expression Subject to Measurement Error
|
|
Author(s):
|
Wenqing He*+
|
|
Companies:
|
University of Western Ontario
|
|
Address:
|
1151 Richmond Street North, London, ON, N6A 5B7, Canada
|
|
Keywords:
|
Accelerated Failure Time Model ; Measurement Error ; Microarray Gene Expression Data ; Survival Analysis ; Variable Selection
|
|
Abstract:
|
Microarray is essentially a measurement tool for measuring biological features such as changes in gene expression levels. Gene expressions could be employed as predictors for patient survival, however, the measurement errors involved in the expression measurement data are often ignored in analysis of microarray data. Efforts are needed to establish statistical method suitable for analyzing microarray data by accounting the error in measuring gene expression. We studied the impact of measurement error on survival relevant gene selection under AFT model setting by applying the regularized weighted least square estimator combined with adaptive LASSO penalty. The simulations show that ignoring measurement error will affect survival relevant gene selection. A SIMEX method is investigated to adjust the measurement error in gene selection.
|
