Discriminating between individual-based models of collective cell motion in a benchmark flow geometry using standardised spatiotemporal patterns

TL;DR

This paper compares five different computational models of collective cell migration within a benchmark flow geometry, analyzing how model assumptions influence behavior and providing guidance on model selection for biological questions.

Contribution

It introduces a systematic comparison of five diverse cell migration models using a standardized flow benchmark, highlighting their differences and applicability.

Findings

Models show distinct flow and deformation patterns.

Model assumptions significantly influence collective behavior.

Recommendations for selecting appropriate models based on research needs.

Abstract

Collectively coordinated cell migration plays a role in tissue embryogenesis, cancer, homeostasis and healing. To study these processes, different cell-based modelling approaches have been developed, ranging from lattice-based cellular automata to lattice-free models that treat cells as point-like particles or extended detailed cell shape contours. In the spirit of what Osborne et al. [PLOS Computational Biology, (2017) 13, 1-34] did for cellular tissue structure simulation models, we here compare five simulation models of collective cell migration, chosen to be representative in increasing order of included detail. They are Vicsek-Gr\'egoire particles, Szab\'o-like particles, self-propelled Voronoi model, cellular Potts model, and multiparticle cells, where each model includes cell motility. We examine how these models compare when applied to the same biological problem, and what differences in behaviour are due to different model assumptions and abstractions. For that purpose, we use a benchmark that discriminates between complex material flow models, and that can be experimentally approached using cell cultures: the flow within a channel around a circular obstacle, that is, the geometry Stokes used in his historical 1851 experiment. For each model we explain how to best implement it; vary cell density, attraction force and alignment interaction; draw the resulting maps of velocity, density and deformation fields; and eventually discuss its respective advantages and limitations. We thus provide a recommendation on how to select a model to answer a given question, and we examine whether models of motile particles and motile cells display similar collective effects.