Catalytic chemical systems are truly multiscale in nature, and have the additional complexity of being composed of heterogeneous materials. Biological systems are also multiscale, and are often of larger dimension. The complexity and interconnected nature of these systems means that modeling purely from first principles is not practical, and optimally designed experimentation is required to capture and predict the behavior of these systems. As with most other multiscale systems, experiments provide information about variables at the coarser or macro scales, while many important variables that are to be inferred and controlled operate at the micro and nano scales. A hierarchy of models that connect across length and time scales is necessary to effectively understand the behavior of these systems. The multiscale models are used to generate optimal experimental designs (input sequences and operating conditions), and the experiments are used to identify and estimate the parameters of the multiscale models.

My current research focus is the optimal model-based design of experiments for catalytic systems, specifically those used for energy generation (e.g., ammonia decomposition for hydrogen production). The energetics of the components of a catalytic reaction mechanism are obtained from first principles quantum calculations (density functional theory) or semi-empirical techniques (UBI-QEP). The effects of catalyst structure are captured using kinetic Monte Carlo models, and microkinetic models are used at the macroscale to connect with experiments. The multiscale models are used with a variety of model based experimental design techniques to identify the maximum number of system parameters. In particular, dynamic information from optimally designed transient experiments is used to maximize system identifiability, and to design and control input sequences and operating conditions.

Another area of my current research is studying the effect of variations at the genetic level on the phenotypic level. Once again, models are used in a hierarchy – stochastic models for genetic transcription, continuum metabolic models and simple cell growth and division models are used along with experimental information at the cellular and protein distribution level. The models are used to understand the mechanisms of regulation in the GAL genetic switch in saccharomyces cerevisae, and the contributions of intrinsic and extrinsic noise to the phenotypic behavior of the system.

The ultimate goal is the effective design and control of these systems based on a combination of multiscale models and experiments – this translates to selecting optimal catalysts and operating conditions (to maximize selectivity and conversion, for example) in catalytic systems. In systems biology, it means controlling the phenotypic response by manipulations at the molecular and genetic level. My future research will focus on attempting to bridge the materials gap and understand, predict and improve the behavior and performance of real catalysts through modeling and experimentation. In systems biology, I will focus my attention on the elucidation of the links between intracellular events (e.g., signaling networks implicated in cancer) and the system or population-level behavior of cells in a tissue or organ. Specifically, I am to study the evolution of cancer progression, which runs the gamut from intracellular events including mutations to tissue invasion and metastasis. Another area of focus is the discrimination between molecular and genetic level mechanisms from population distribution data, which is the normal type of experimental data available at the protein and cellular level. This is a difficult problem for which no standard methods exist at present, but is of enormous importance in systems biology.