Biomarker identification for early tumor detection aided by bioinformatics gene expression analysis

Abstract

Early detection of cancer is key to prompt treatment and patient's long term survival. Designing noninvasive and simple methodologies that can be used in large scale screening for early tumor detection is an important research area that could have profound impact on cancer prevention and treatment. Taking advantage of available gene expression databases, we have performed extensive analysis on differentially expressed (DE) genes and alternatively spliced (AS) forms of genes in cancer tissues. In this study we detected genes that are potentially up-regulated in tumor tissues comparing to normal tissues. We also aligned the 8 million human EST sequences in dbEST with human genomic sequences, and found several hundreds AS variants that have a potential to be specific to or overexpressed in tumor tissues comparing to normal tissues. Our results showed the possibility of using gene expression databases to detect potential "cancer genes", and to use those DS genes and AS variants of genes as markers for early cancer detection. © 2008 IEEE.