Similarity of Multivariate Methods to Establish Microarray Quality Control Standards

ABSTRACT

This article describes a multivariate method based on logistic regression that distinguishes between usable and low-quality digital image spots associated with DNA microarray data analysis. To create a response variable describing spot quality, the authors analyzed the similarity among decisions that were made with control charts applied to quality parameters established in the gene expression measurement literature.

KEYWORDS:

• DNA microarrays
• gene expression data
• Hotelling's T ²
• logistic regression
• multivariate methods
• principal components
• quality control
• regression adjustment
• similarity

ACKNOWLEDGMENT

Support has been provided by the Arkansas Biosciences Institute, the major research component of the Arkansas Tobacco Settlement Proceeds Act of 2000.

Notes

Shading indicates maximum likelihoods..

Shading indicates combined quality parameter values assumed to indicate unusable spots.