ABSTRACT
Rapid developments and applications of various genomics and proteomics platforms in recent years have revolutionized biomedical research. Instead of focusing on one or a few candidate genes, researchers now routinely collect the expression information of all transcripts in the genome, study their coexpression patterns, and correlate expression profiles with disease phenotypes and clinical outcomes. One of the most cited papers is Golub et al. (1999), where the authors demonstrated that gene expression profiles can be used to reliably define leukemia subtypes (Golub et al. 1999). This approach has been applied to essentially all cancer types where gene expression levels from a small number of genes can be used to characterize tumor heterogeneity. Earlier platforms were based on gene expression microarrays, whereas next-generation sequencing is more commonly used now to measure gene expressions as well as other molecular phenotypes, for example, methylation to discover and define cancer subtypes. In these cases, the genes selected that represent cancer subtypes can be regarded signature biomarkers.
