Collaborative Research: DMS/NIGMS 1: Identifiability investigation of Multi-scale Models of Infectious Diseases

The emergence and re-emergence of pathogens and their impact on society has reinforced the need for integration and synergy across scientific fields and biological scales in order to advance understanding, predicting, and responding to pathogen spread. Multi-scale mathematical models that consider the timing and length of individual infections when modeling transmission into the population can aid recommendations for optimal interventions. One shortcoming when evaluating data using multi-scale models comes from data scarcity in the expansion stages of the infection and transmission, the differences in data magnitude and frequency at each scale, together with the complexity of the models considered. To determine the source of combined biases in parameter estimation, we will use a combined empirical-theoretical approach for investigating structural and practical parameter identifiability of multi-scale models of infectious diseases that may inform optimal experimental design. The proposed research will facilitate a better understanding of the sources of uncertainty when fitting multi-scale models to multi-scale infectious disease data, with a focus on Usutu and SARS-CoV-2 viruses. By combining empirical and theoretical approaches we aim to determine structural and practical parameter identifiability of multi-scale models, to inform optimal experimental design, and to improve our ability to make predictions and suggest interventions. Our proposal will focus on three major mathematical challenges: (1) Developing methods for improving practical identifiability in within-host systems; (2) Use experimental data to inform development of transmission models; (3) Build a quantitative framework to predict parameter identifiability in multi-scale systems. The overarching goal of the proposed work is to integrate multi-scale mathematical model development and statistical models for data fitting with collection of longitudinal virus titers and probability of transmission data in order to decrease uncertainty and improve results reproducibility. This will ultimately improve our understanding of infection disease transmission and persistence.