|
Activity Number:
|
242
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #302216 |
|
Title:
|
Rediscovering the Power of Well-Planned Comparisons: Normalization and Analysis of cDNA Microarray Using Linear Combinations
|
|
Author(s):
|
Liping Huang*+ and Naoki Miura and Michael Mienaltowski and James MacLeod and Arnold Stromberg and Arne Bathke and Constance Wood
|
|
Companies:
|
University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky
|
|
Address:
|
Departmen of Statistics, Lexington, KY, 40508,
|
|
Keywords:
|
cDNA array ; M-A plot ; linear combinations ; factorial design
|
|
Abstract:
|
A recurring issue in the analysis of cDNA microarray data is normalization, the purpose of which is to adjust for systematic variation. The important source of systematic variation is dye bias. Planned linear combinations eliminates dye and chip effects by negation and then constructs a sequence of comparisons to identify probe sets with differential expression levels. We exemplified our methods in two datasets. The first dataset was a comparison between normal cartilage and repair tissue and the second one is a 2x2 factorial treatment design with a replicated split block experimental design. Under the appropriate experimental design and with studies lacking strong biological assumptions, the linear comparison approach, when used judiciously, constitutes a powerful and transparent alternative to an intensity dependent normalization LOWESS method and to mixed model approach.
|
