Channeling the Data Flood: Handling Large-Scale Biomolecular Measurements in Silico

The cells of the human body each contain thousands of molecular species,the dynamic interactions of which constitute the biomolecular system underlying cellular function. Many thousands of expressed genes, proteins, and metabolites can be measured simultaneously in a tissue or blood sample, making a scan of most biomolecules possible. This wealth of information causes the bottleneck in biomedical research to shift from making measurements to data integration and analysis. Management of biomolecular information requires efficient data storage and retrieval from integrated databases. Subsequent data analysis has three components: 1) identification of broad molecular "fingerprints" useful for early disease diagnosis and treatment selection; 2) identification of the relatively few molecular signals which change significantly above the high noise level formed by biological variation amongst individuals; and 3) development of a mechanistic understanding of the system by capturing its characteristics in computational models. At present, extensive parts of the biomolecular system are still uncharted. Representation of the entire system in silico presents a great challenge, made more challenging still by great interindividual variation at the molecular level. The prospect of much better control of human disease makes this gigantic enterprise worthwhile.