Resilience in hierarchical modular networks: How multi-scale biosystems cope with damage

• Applicant: Privatdozent Paolo Moretti, Ph.D.
• Subject Area: Statistical Physics, Nonlinear Dynamics, Complex Systems, Soft and Fluid Matter, Biological Physics
• Term: from 2017 to 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 394689530

Resilience is a central issue in the study of biological networks. Living systems often epitomize the ability to cope with perturbation and damage, and their network structure is believed to be at the origin of such ability. The concept of network resilience was introduced in order to investigate and model the fault tolerance of techno-social complex systems such as power grids or web infrastructures. It was found in this context that scale-free network structures, characterized by a backbone of highly connected hubs, showed an unprecedented resilience in terms of an outstanding ability to maintain system-wide connectivity (existence of a giant connected cluster) despite widespread random failures of links or nodes. Recent advances in bio-imaging techniques, however, have shown that many biological networks (e.g., brain networks, gene regulatory networks, collagen fiber networks) depart significantly from the scale-free network paradigm, exhibiting hierarchical modular network (HMN) structures instead. However, a detailed study of resilience of HMNs is currently lacking. This project aims at filling this gap by providing an extensive numerical study of aspects of resilience in HMNs that may be of relevance in biosystems modeling. Furthermore, biological networks are multi-functional and capable of simultaneously fulfilling differentiated tasks, a capability which may be insufficiently represented by the mere existence of a giant connected cluster. Hence, more sophisticated indicators of resilience, envisaged as the capability of sustaining complex dynamics on the network, may be required. To address these dual aims, we use concepts of network theory and apply them to numerical models of random failure as well as of activity-induced damage in HMNs. We probe how damage affects the structural morphology of HMNs and associated parameters, in particular the network dimension, and what implications such changes have on dynamical processes operating on the networks. We apply this methodology both to computer generated networks and to experimental data of network connectivity in specific examples of resilient biosystems, namely brain networks, with the objective of developing efficient analysis tools that can be deployed in more general contexts to characterize resilience of biological networks.

• DFG Programme: Research Grants