Challenges of Complexity and Integration in Quantitative Systems Biotechnology

Presented at NSF Workshop on Quantitative Systems Biotechnology, September 13-14, 2000It is widely acknowledged that the enormous amount of data produced by the human and other genome projects requires new mathematical and computational strategies of analysis and integration. Classified very broadly, these strategies may either address large-scale databases and networks, or they may focus on the intricate details of smaller regulatory systems. Eventually, the two approaches must converge, but our current methodologies may not be sufficient for this to happen now. At this point, typical large-scale approaches are designed to detect relationships among genes or between gene expression and function. These approaches make extensive use of the ability of computers to address combinatorial problems with high efficiency. Although very successful in many respects, these approaches alone are not sufficient. They must be complemented with detailed algebraic and numerical analyses that aim at discovering and explaining the design principles behind natural systems and at integrating diverse pieces of information within and between levels of biological organization. It has been shown that such analyses can help explain why a gene circuit or a metabolic pathway is regulated in particular way and not in another, theoretically possible fashion. Detailed smaller-scale analyses can lead to a rationale for why genes of the same pathway may be over-expressed at drastically different rates, when the organism is exposed to a stimulus. They may provide reasons for why the artificial over-expression of genes in a biotechnological setting does not necessarily result in the desired and expected increase in product yield. The presentation primarily discusses the challenges of integration in complex systems and briefly mentions a mathematical approach, based on power-law approximation, that has been helpful in dealing with some of these challenges