Understanding the Emergence of Topological Defects in Cell Monolayers
Oral-In-person · Withdrawn
Abstract
In biological tissues, individual cells might be thought of as simple computational agents. In cellular monolayers specifically, cells process mechanical and biochemical inputs and make basic, non-determinate “decisions” which result in behaviors such as polarity alignment and directed migration. At the collective scale, the transmission of basic information between cells and local decisions generate emergent patterns, including nematic alignment, topological defects, and flow domains reminiscent of liquid crystals.
To investigate the origins of these characteristics we develop a morphodynamic network model (MNM) where a tissue monolayer is represented as a dynamic network of nodes (cell centers) and edges (cell–cell contacts). Unlike traditional vertex or Voronoi-based models, MNM permits a broader range of cell geometries and dynamics; a specific characteristic this model allows is migration along an elongated cell’s major axis – a behavior which appears to be tightly coupled to nematic ordering.
The MNM evolves via two decoupled processes: force-driven node motion capturing cellular mechanics, and stochastic T1 rearrangements simulating neighbor exchanges. MNM reveals that migratory ±1/2 topological defects emerge spontaneously under rearrangement-dominated conditions, reflecting an underlying interplay between cellular motility, contractility, and interfacial tension.
To study how basic computational properties of cells lead to these emergent behaviors and structures, we develop a machine learning-based pipeline that analyzes simulation images to detect, classify, and quantify defect structures. This enables a data-driven link between microscopic cellular rules and macroscopic defect dynamics.
This work offers a view of tissues as information-driven active materials, where structure and function arise through decentralized, cell-level computation and mechanical feedback. This perspective opens new avenues for modeling living systems as soft active matter with embedded information-processing capabilities.
To investigate the origins of these characteristics we develop a morphodynamic network model (MNM) where a tissue monolayer is represented as a dynamic network of nodes (cell centers) and edges (cell–cell contacts). Unlike traditional vertex or Voronoi-based models, MNM permits a broader range of cell geometries and dynamics; a specific characteristic this model allows is migration along an elongated cell’s major axis – a behavior which appears to be tightly coupled to nematic ordering.
The MNM evolves via two decoupled processes: force-driven node motion capturing cellular mechanics, and stochastic T1 rearrangements simulating neighbor exchanges. MNM reveals that migratory ±1/2 topological defects emerge spontaneously under rearrangement-dominated conditions, reflecting an underlying interplay between cellular motility, contractility, and interfacial tension.
To study how basic computational properties of cells lead to these emergent behaviors and structures, we develop a machine learning-based pipeline that analyzes simulation images to detect, classify, and quantify defect structures. This enables a data-driven link between microscopic cellular rules and macroscopic defect dynamics.
This work offers a view of tissues as information-driven active materials, where structure and function arise through decentralized, cell-level computation and mechanical feedback. This perspective opens new avenues for modeling living systems as soft active matter with embedded information-processing capabilities.
–
Presenters
-
Aidan Smith
- University of Colorado, Boulder