Reverse engineering the inflammatory “clock”: from computational modeling to rational resetting

Abstract

Properly-regulated inflammation is central to homeostasis. Traumatic injury, hemorrhagic shock, septic shock, and other injury-related processes such as wound healing are associated with dysregulated inflammation. Like many biological processes, inflammation is a dynamic, complex system whose function, like that of an analog clock, cannot be discerned simply from a laundry list of its parts (data). The advent of multiplexed platforms for gathering biological data, while providing an unprecedented level of detailed information about the inflammatory response, has paradoxically also proven to be overwhelming. This problem is especially acute when the datasets involve time courses, since typical statistical analyses and data-driven modeling are geared towards single time points. Various groups have addressed this problem using dynamic approaches to data-driven and mechanistic computational modeling. These modeling tools can be thought of as the “gears” and “hands” of the “clock,” and have led to insights regarding principal drivers, dynamic networks, feedbacks, and regulatory switches that characterize and perhaps regulate the inflammatory response. In parallel, mechanistic computational models have given an abstracted sense of how the inflammatory “clock” works, leading to in silico models of critically ill individuals and populations. Integrating data-driven and mechanistic modeling may point the way to a rational “resetting” of inflammation via model-driven precision medicine.

Reverse engineering morphogenesis through Bayesian optimization of physics-based models

“…Reverse engineering biological systems requires mathematical tools to infer interactions and mechanisms of a given process to enable forward engineering applications 1 , 2 . For instance, a detailed understanding of the pathways involved during a wound-healing process can help design new treatment plans or drugs 3 , 4 . The calibration of computational models of biological systems is challenging due to the large number of interactions whose mathematical description encompasses a very large parameter space.…”