Title: Simulation Based Inference Design of Experiment for Mechanistic Acyclic Networks (sbiDOE MAN)
Abstract: Biological signaling pathways, such as the bone morphogenetic protein (BMP), Wnt, and JAK-STAT, can be modeled as systems of ordinary differential equations (ODEs) based on mass action kinetics. One common problem biologists face when studying these pathways is to determine the unknown parameters of the pathway that represent physically relevant features, such as binding affinity of ligands to receptors. However, the number of combinations and concentrations to test to determine latent parameters, such as ligand-receptor binding affinity, can become prohibitively large to evaluate using brute force screening methods. Bayesian optimal experimental design methods can be used with prescribed mathematical models of signaling pathways’ activity to determine optimal experiments to perform to reduce uncertainty in parameter estimates for prescribed ODE models. An example of sbiDOE MAN will be demonstrated for a simple mathematical model of the BMP signaling pathway.