Computational Biology offers a valuable complement to complex experimental approaches to study infectious diseases like HIV and tuberculosis. Computational Biology uses mathematical, statistical, and machine learning models to create ‘digital twins’ of complex biological interactions. The computational representations can easily and cost-effectively quantify relative contributions of each biological mechanism to the overall system dynamics, identifying weak links that can be exploited to subvert the pathogen. However, development of these digital twins typically rely on ‘real world’ assay development first. In this project, I propose that the digital twins need to be developed in parallel with the real-world assays. By continually training the digital twins throughout the development of the real-world assay, we will more closely represent all aspects of the biology, not just the limited set of ‘final’ data. Through this approach, we could generate a comprehensive ‘virtual lab’ that can closely mimic the process of developing new assays, thereby broadening the impact of computational models beyond individual assays.
This work will focus on producing proof of concept of this approach for TB as a test case, using both traditional mechanistic computational models as well as artificial intelligence (AI) (or a combination of the two) to accomplish this goal.