|
Activity Number:
|
374
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306970 |
|
Title:
|
Array Spatial Variability and Normalization Techniques for Microarray Gene Expression Signals
|
|
Author(s):
|
Samir Lababidi*+ and Daya Ranamukhaarachchi
|
|
Companies:
|
U.S. Food and Drug Administration and U.S. Food and Drug Administration
|
|
Address:
|
1350 Piccard Drive, Rockville, MD, 20850,
|
|
Keywords:
|
microarrays ; gene expression ; spatial variability ; normalization
|
|
Abstract:
|
We previously showed that microarray gene expression data can have substantial variability based on spatial arrangement of genes on the chip and that such data, without correction for array spatial variability, may not be reliable enough to be used in disease diagnosis and prognosis. In this talk, we will present a chip design to study array spatial variability for two-color arrays across the chip surface and different times of hybridization while accounting for possible dye effect. To explore these sources of variability, appropriate statistical models were developed and multiple testing procedures implemented to take into account the high-dimensionality of multiple genes. These models can be used to obtain a method for normalization of gene expression in which measurements for these genes would be more reliable.
|
