Statistical Analysis Can Fail to Reveal Underlying True Biological Mechanism: A Demonstration of Expression Profile Generation

In the life science, data show an exponential increase trend with the development of biotechnology. Statistical analysis is a powerful tool to find useful information in tremendous amount of data. However, biological systems always contain serious errors where data-driven biostatistics may meet barriers. Different and complementary approaches are needed. Endogenous network theory is a mechanism-driven method which can quantitatively analyze regulation networks. Steady states expression profiles and potential landscape of a dynamic network could be found by this theory. Simulations of stochastic effects indicate that in systems with large noise, data-driven biostatistics can fail to reveal the true biological mechanism.