Agent-based and in silico models of cell behaviour
In silico models run a biological idea as a program and compare the output with experiment. Agent-based versions treat cells, or patches of cell membrane, as autonomous units with local rules. Budnikova and colleagues argue that agent-based models hold promise for studying emergent behaviour in regeneration signalling, when multi-scale or multi-cellular systems are supported.1 Predictions of the memAgent-spring model have been validated in vivo, according to Zakirov and colleagues.5
- Earliest held
- 2009, Fernando CT, Liekens AM…
- Most discussed in
- Active perception during angiogenesis…, 2021
- In the library
- 148 passages in 38 works
- Rewritten
- 2026-10-26
Cells as agents
Budnikova and colleagues (2014) described a plan to combine agent-based modelling with an evolutionary search engine to find and validate computational models of regeneration. They argued that agent-based modelling holds promise for studying the emergent behaviour of signalling networks, when multi-scale or multi-cellular systems are supported.1 Their platform, CellSim, uses autonomous cells as the central agents. These carry many of the biological primitives needed to simulate living systems.2 The authors noted that building, testing and validating a complex model by hand is hard. The search engine was added to automate part of that work.2
Membrane agents and springs
Zakirov and colleagues (2021) used the memAgent-spring model of angiogenesis. A cell is drawn as a surface of many small membrane agents joined by springs that follow Hooke's Law. The springs stand in for the tension of the actin cortex under the cell membrane.6 The agents move in continuous space but are snapped to a lattice so that discrete local rules can apply.7 An earlier, simpler memAgent model had predicted that filopodia and Vegf gradients alter selection speed. Zakirov and colleagues report that the spring version recapitulates that prediction.5
Testing against animals
Lobo, Lobikin and Levin (2017) worked with a model that had been compared with Xenopus experiments. Those experiments had shown two extreme outcomes, normal or high conversion. The authors defined a distance metric to search for simulated phenotypes between these extremes.3 Kriegman, Blackiston, Levin and Bongard (2020) compared simulated and realised designed organisms. They removed cilia in vivo and left them out of the simulation. They compared trajectories upright and inverted. For at least one design, the data suggest the desired behaviour transferred when upright but not when inverted.4
Recent frameworks
Pio-Lopez, Bischof, LaPalme and Levin (2023) coded their model of goals in Python. They used the agent-based framework Mesa and the MultiNEAT package, and released it publicly.8 Desoyer and colleagues (2026) examine whether cancer is a bioelectric disease. They describe computational modelling and emerging digital twin frameworks for simulating membrane voltage dynamics. They say these frameworks allow hypothesis testing in silico. They call this a step towards predictive, multiscale modelling of bioelectric regulation in cancer.9
Agent-based modeling holds promise for studying the emergent behavior and complex interactions between signaling networks involved in directing regeneration
Budnikova M, Habig JW, Lobo D, Cornia N, Levin M, Andersen T, 2014 · Design of a flexible component gathering algorithm for converting cell-based… · open at passage 7we are using a modeling platform (CellSim) where the central agents are autonomous cells containing many of the biological primitives necessary for simulating living systems
Budnikova M, Habig JW, Lobo D, Cornia N, Levin M, Andersen T, 2014 · Design of a flexible component gathering algorithm for converting cell-based… · open at passage 7To discover phenotypes different from the two extreme levels of conversion (normal or high) previously observed in silico with the model and in vivo with Xenopus experiments
Lobo D, Lobikin M, Levin M, 2017 · Discovering novel phenotypes with automatically inferred dynamic models: a… · open at passage 21Trajectories of deciliated designs are compared in silico and in vivo, in two orientations (upright and inverted 180° about the transverse plane)
Kriegman S, Blackiston D, Levin M, Bongard J, 2020 · A scalable pipeline for designing reconfigurable organisms · open at passage 13Early simulation predictions with the memAgent model [24] predicted that filopodia and Vegf gradients alter selection speed.
Zakirov B, Charalambous G, Thuret R, Aspalter IM…, 2021 · Active perception during angiogenesis: filopodia speed up Notch selection of… · open at passage 20connected by springs following Hooke's Law (which confer tension to the changing cell shape, akin to the actin cortex beneath a cells membrane)
Zakirov B, Charalambous G, Thuret R, Aspalter IM…, 2021 · Active perception during angiogenesis: filopodia speed up Notch selection of… · open at passage 78The memAgents move in continuous space but are snapped to a gridded lattice to implement discrete local rules of agent interactions
Zakirov B, Charalambous G, Thuret R, Aspalter IM…, 2021 · Active perception during angiogenesis: filopodia speed up Notch selection of… · open at passage 78The whole system has been coded in Python programming language, using the agent-based modelling framework Mesa [64] and the MultiNEAT package.
Pio-Lopez L, Bischof J, LaPalme JV, Levin M, 2023 · The scaling of goals from cellular to anatomical homeostasis: an evolutionary… · open at passage 30These frameworks enable hypothesis testing in silico and represent an important step towards predictive, multiscale modelling of bioelectric regulation in cancer.
Desoyer C, Šupe R, Ghorbanpour S, Goswami D, Khalaj M…, 2026 · Cancer: A bioelectric disease? · open at passage 92