BIMSA >
Deep Learning at the Frontiers of Science
Lattice Models for Coarse-Grained Representation of Dynamic Biological Systems
Lattice Models for Coarse-Grained Representation of Dynamic Biological Systems
Organizers
Speaker
Yusong Ye
Time
Thursday, September 18, 2025 10:30 AM - 11:30 AM
Venue
A3-1-101
Online
Zoom 293 812 9202
(BIMSA)
Abstract
Agent-based models have gained significant attention in systems biology due to their ability to accurately capture key features of biological systems while maintaining computational efficiency. These models offer a flexible framework that represents both macro- and micro-level physical properties but also reduced the computational complexity. This presentation focuses on lattice models—a wide-used agent-based approach: The most famous coarse-grained statical physic models - Ising model is a typical lattice model. We will discuss their characteristics, methodologies, and relevant research challenges.
A major advantage of lattice models is their ability to simulate different scales of biological activity with relative simplicity. They are particularly suitable for studying the growth processes in cell colonies. For example, we applied lattice model to simulate biofilm growth dynamics—a type of unstructured multicellular community—to examine its growth patterns with experimental results. Our simulations clearly show that the growth can depend on gel structure and nutrient distribution. Additionally, biofilm healing patterns and capacity are influenced by nutrient availability.
In summary, lattice models effectively replicate experimental observations and enable useful predictions. Their potential extends beyond microbial communities but also in cancer tumor research.
A major advantage of lattice models is their ability to simulate different scales of biological activity with relative simplicity. They are particularly suitable for studying the growth processes in cell colonies. For example, we applied lattice model to simulate biofilm growth dynamics—a type of unstructured multicellular community—to examine its growth patterns with experimental results. Our simulations clearly show that the growth can depend on gel structure and nutrient distribution. Additionally, biofilm healing patterns and capacity are influenced by nutrient availability.
In summary, lattice models effectively replicate experimental observations and enable useful predictions. Their potential extends beyond microbial communities but also in cancer tumor research.
Speaker Intro
Dr. Yusong Ye earned his Ph.D. in Mathematics from Beihang University under Professor Yang Zhuoqin and completed postdoctoral research at Friedrich-Alexander University Erlangen–Nuremberg (FAU) under Professor Vasily Zaburdaev before becoming a Lecturer at Beijing University of Petrochemical Technology. His research uses mathematical modeling to study biological systems, integrating physical and biological data to understand processes like tumor growth and, more recently, the dynamics of toxic proteins in neurodegenerative diseases.