Abstract: A mathematical model is constructed based on the current knowledge about a biological system to predict its dynamics and test hypotheses of the system of interest. However, predictions from such models are often subject to high uncertainty since many biological systems are only partially understood. To address such uncertainties, I proposed a hybrid modeling approach to improve model accuracy by combining machine learning and the first-principle modeling. Specifically, model correction terms are learned from discrepancy between model predictions and measurements, and these terms are added to the first-principle model to enhance the prediction accuracy. Once these correction terms are learned from the data, an artificial neural network (ANN) model is developed to find an empirical relation between the model and the correction terms so that the developed ANN can be used to posses improved predictive capabilities even in new operating conditions (i.e., generalizability). The final hybrid model is then constructed by coupling the first-principle model with the developed ANN. As case studies, the proposed approach was implemented to develop hybrid models using in silico and in vitro datasets to describe two different systems more accurately when the available mechanistic models were not complete.
Bio: Dr. Dongheon Lee is a postdoctoral researcher at Duke University working with Professor Lingchong You. He received B.S. in Chemical and Biomolecular Engineering from Rice University in 2015 and Ph.D. in Chemical Engineering from Texas A&M University in 2020. His current research focus is to engineer liquid-liquid phase separation (LLPS) to control complex reaction networks in cells. Specifically, he is interested in using LLPS for engineering metabolic pathways as well as designing synthetic genetic circuit. During his graduate study with Professor Joseph S. Kwon, he leveraged his training in process systems engineering to propose methodologies to identify and calibrate mathematical models for describing complex biological systems such as intracellular signaling pathway and lectin-glycan binding kinetics. Dongheon has published 11 papers relating to his work, in addition to six peer-reviewed conference proceedings.