Integrating systems biology data repositories to improve cis-regulatory motif characterization and discovery: An elegant solution to an old difficult problem.

摘要

In the post-genomic age, algorithmic techniques have been developed to determine the presence and location of genes in genome sequences. Successful algorithmic techniques for predicting the structure and function of the modules regulating these genes have yet to be presented in a unified theory. In the past, scientists have explored the diversity and evolution of regulation control mechanisms through genome wide snapshots of the expression profile (e.g. micro-arrays) or with detailed studies of regulatory control mechanisms involving a few genes. These approaches were important first steps, but are problematic because they did not consider system wide activity of regulatory proteins. In recent years, detailed regulatory networks have been described in bacteria. These genome wide databases of regulatory network interactions allow development of a more integrative genome wide approach to describing regulatory control. This thesis addresses three obstacles to discovering the structure of regulatory interaction networks: (1) it allows ranking of mathematical methods for modeling transcription factor-DNA interactions based upon predictive ability; (2) utilizing the regulatory regions from multiple organisms, it increases prediction efficiency of the modeling methods, and (3) it introduces a new paradigm for binding site prediction based on regulatory interaction mechanisms of transcription factors with RNA polymerase. This new integrative approach offers entirely new avenues for future research into the regulatory cascades in the cell. With a complete regulatory map in hand, science may unlock cellular behavior and control.