Home | Business News | Browse by Publication | J | Journal of the American Statistical Association

Semilinear high-dimensional model for normalization of microarray data: a theoretical analysis and partial consistency; Comment.

Article Excerpt
1. INTRODUCTION

I would like to begin by congratulating Fan, Peng, and Huang on their innovative approach to normalizing spotted cDNA microarray data. Although numerous articles have been written on this topic, the use of duplicate spots in the normalization process has rarely been discussed. Given the apparent within-slide spatial effects (Smyth, Yang, and Speed 2003; Balazsi, Kay, Barabasi, and Oltvai 2003), these spots have the potential to be very helpful not only in the normalization process itself, but also in assessing the effectiveness of a normalization procedure. Unfortunately, the focus of the article limits the discussion of this approach and leaves several questions regarding the implementation and effectiveness unanswered. It is on these questions that I want to focus here, in the hope of prompting further investigation into the use of these spots.

2. MULTIPLE VERSUS SINGLE SPOT NORMALIZATION

Much of the popularity of microarray analysis stems from the fact that scientists can view the simultaneous behavior of genes affected by a stimulus at the total genome level. Because duplicate spots take up space on the slide, there is a general reluctance to include them unless there are no other probes of interest (exceptions would be housekeeping or control probes). In addition, having duplicate spots often means that additional genetic material (i.e., labeled cDNA) must...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!