A method for linking microarray data to database information

Abstract

Microarray technology is a new method of examining the whole genome expression profile. There are over 2,000 manuscripts in total published on microarray data analysis. There are also many database resources for genomic information. Although the human genoine is largely known, the degree to which each gene is expressed is not known. Numerous authors have addressed this problem using microarray or database information. Here, we develop an approach that links the gene expression microarray data with the available genomic database information spatially. This provide and alternative method to clustering tools in examining differential gene expression. Microarray data were pre-processed and normalized so that the data are independent of the technology. Out of 9,600 genes from the original microarray slide, there are 8474 normalized genes for subsequent linkage with the genomic database. The choice of genomic database is Gene Ontology (GO). The Genomic database information was represented in a 2-dimensional map using Correspondence Analysis (CA). The normalized microarray data were then linked with the genomic database information using response surface methodology.