Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 28: Poster: Machine Learning, Data Science

DY 28.7: Poster

Mittwoch, 19. März 2025, 15:00–18:00, P4

Parameter estimation and Bayesian comparison of Langevin models describing cell motility — •Yusuke Kato^1,2, Jan Albrecht¹, Hiroshi Kori², Robert Großmann¹, and Carsten Beta¹ — ¹Institute of Physics and Astronomy, University of Potsdam, Germany — ²Department of Complexity Science and Engineering, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan

In nature, motile bacteria and eukaryotic cells exhibit spontaneous movement. This cell motility plays an essential role in both maintaining homeostasis (such as the migration of immune cells to a wound site) and the pathogenesis of certain diseases (like the aggregation of cancer cells to other organs in metastasis). Various SDE-based Langevin models have been proposed to describe cell motility.

In this study, we adopt a Bayesian approach using the likelihood approximation technique introduced in Ref. [arxiv:2411.08692] to estimate parameters of tentative first- and second-order models. In order to compare the different models, we develop a framework that ranks them using Bayesian model comparison. We test and benchmark the approach using synthetic data and subsequently apply it to time-series data of ameboid cells in order to find the best model for their ameboid motility.

Keywords: cell mortility; parameter estimation; Bayesian model comparison; Langevin equation