Title: Bounded System Identification for Biological Systems

Speaker: Dr. Cranos Williams, North Carolina State University

Date/Time: Monday, April 12, 2010, 10:00 am      

Location: CSRI Building/Room 147 (Sandia NM)

Brief Abstract: Current modeling approaches produce estimates that do not reflect the uncertainty associated with biological processes that stem from biological variation, uncertain experimental measurements, and unknown or under-modeled aspects of the process.  Models that do not adequately reflect this uncertainty do not capture the full possible range of dynamic activity, which results in models with reduced predictive ability.  This can have an adverse effect on predicting the effects of treatments and other modifications on these processes.  Assessing the full range of possible activity gives the modelers and the experimentalist a realistic assessment of what the biological process ‘can’ do. Analytical and computational approaches that are capable of quantifying this uncertainty are needed.  In our research, we develop and refine bounded approaches for estimating the range of component concentrations and kinetics parameters that are consistent with the uncertainty associated with biochemical pathways.  These methods are based on Interval Analysis and are modified to address the hurdles typically associated with modeling biological systems.  These methods exploit the strengths of bounded approaches for improving the predictive power of biological models.  Future applications include design of the specifications for wet-lab experiments and model reduction.

CSRI POC: Elebeoba May, (505) 844-9933