Data Mining for Biomarker Discovery: The Time is Ripe

摘要

Identifying molecular biomarkers has become an essential topic of study in biological data mining, bioinformatics, and translational medicine today. Although the exact definition for biomarkers is still up for debate, it is commonly used to refer to the presence, variation, or modifications of biomolecular entities such as genes, proteins, and metabolites that may be used to assess different phenotypic states of cells or organisms. Advances in DNA sequencing, functional genomics, genome-wide association studies, proteomics, metabolomics, and network biology have enabled biomedical researchers to compare quantitatively changes of multiple genes, proteins, metabolites, and compounds between case and control biological samples. The ongoing influx of these study data, while still inconsistent when performed in different laboratories, has nonetheless created new and exciting opportunities for data mining researchers in the post-genome era. How to identify candidate molecular biomarkers that has good overall performing in assessing disease risks, detecting disease early, dissecting disease subtypes towards tailored treatment selection, and improving outcomes of disease treatment - all have become critical questions for applying bioinformatics to future predictive and personalized medicine.