An associative analysis of gene expression array data

• DOI: 10.1093/bioinformatics/19.2.204
• Publication Date: 2003-01-21T20:47:16Z
• Authors: Igor Dozmorov; Michael Centola
• Keywords: Associative property; Expression (computer science)

Abstract Motivation: We face the absence of optimized standards to guide normalization, comparative analysis, and interpretation data sets. One aspect this is that current methods statistical analysis do not adequately utilize information inherent in large sets generated a microarray experiment require tradeoff between detection sensitivity specificity. Results: present multistep procedure for mRNA expression obtained from cDNA array methods. To identify classify differentially expressed genes, results standard paired t-test normalized are compared with those novel method, denoted an associative analysis. This method associates experimental gene expressions presented as residuals regression against control averaged common standard—the family similarly computed low variability genes derived experiments. By associating changes given equally group, utilizes experiments increase both specificity sensitivity. The overall illustrated by tabulation whose differs significantly Snell dwarf mice (dw/dw) their phenotypically normal littermates (dw/+, +/+). Of 2352 examined only 450–500 were above background levels observed nonexpressed these 120 established at significance level excludes appearance false positive determinations. Contact: [e-mail] * whom correspondence should be addressed.