Molecular profiling approaches for identifying novel biomarkers.

An unprecedented interest in biomarker development has arisen from the increasing use of genomic information and high-throughput technologies in the field of drug development. Monitoring global cellular responses to perturbation due to disease, drug treatment or toxicity is achieved using molecular profiling methods such as DNA microarrays, proteomics and metabonomics. Unique fingerprints composed of molecular changes are captured and subjected to interpretation with the goal of class discovery, comparison or prediction. Each fingerprint reflects a cumulative response of complex molecular interactions, and if these interactions can be significantly correlated to an end point, the molecular fingerprint may be qualified as a predictive biomarker. Furthermore, in cases where the predictive power of any single response or set of responses is statistically significant, a molecular fingerprint can provide novel information related to the underlying disease biology or mechanism of toxicity. There is an acute need for effective biomarkers in every phase of drug development, from discovery, to preclinical studies, through to clinical trials. The context in which these molecular biomarkers are used will depend upon the nature of the biological problem being addressed. This review will summarise experimental and computational efforts in the field of molecular profiling and discuss the significant challenges in interpreting molecular profiling data and qualifying novel transcriptional biomarkers.