Mechanistic Compartment Models of Cellular and Viral Dynamics

45 concepts41 papers

When trying to understand how a disease progresses or how a treatment takes effect, statistical models often look for correlations—matching a genetic marker to a survival rate or a drug dosage to a viral drop. However, correlations do not explain why a system behaves the way it does, nor can they easily extrapolate beyond the data used to train them.

To build a deeper predictive capability, researchers use mechanistic compartment modeling. As argued by Thiébaut et al., this approach shifts the focus from purely statistical patterns to explicit biological "rules" formalized through differential equations. By dividing a complex biological system into discrete pools (compartments) and defining the mathematical rates at which entities move between them (fluxes), these models capture the underlying causal machinery of life, infection, and treatment.

The Anatomy of a Compartment Model

The core intuition behind a compartment model is that a population—whether of cells, virions, or bacteria—is rarely homogeneous. Instead, individuals exist in distinct functional states.

To model this, researchers define a state-space where each compartment represents a specific biological condition:

  • Proliferating vs. Quiescent: To understand how immune therapies restore T-cell counts in HIV patients, Thiébaut et al. constructed ordinary differential equations (ODEs) that explicitly track the transition of cells between active division and resting states.
  • Susceptible vs. Dormant vs. Dead: In studying why oral biofilms resist disinfectants, Shen et al. used coupled ODEs to show that a small subpopulation of "persister" cells shifts into a dormant state during treatment, allowing the bacterial community to fully recover once the stressor is removed.
  • Phenotypic Plasticity: Rather than assuming drug resistance is purely a result of selecting pre-existing mutant cells, Chapman et al. designed a state-space model with a dedicated sink state for cell death. This allowed them to mathematically decouple population changes driven by selective killing from those driven by active "shape-shifting" into drug-resistant states.

By defining these states, the model moves beyond tracking total numbers; it tracks the flux—the rates of birth, death, and phenotypic transformation—governed by environmental variables or therapeutic interventions.

Modeling Hierarchies and Spatial Migration

Biological systems are rarely static mixtures; they are often organized into strict developmental hierarchies or distributed across physical space. Compartment models handle this complexity by chaining differential equations together sequentially.

In hierarchical tissues like blood or skin, cells continuously differentiate from rare stem cells into massive populations of mature effector cells. Werner et al. utilized deterministic population dynamics to track mutant cells across these multi-stage hierarchies. Their analytical solutions proved a fundamental tissue property: while mutations in self-renewing stem cells create permanent mutant clones, mutations in downstream progenitor cells are transient and eventually "wash out" of the system.

[Stem Cells] ---> [Progenitor Stage 1] ---> [Progenitor Stage 2] ---> [Terminal Cells]
     |                     |                         |                      |
  (Permanent)          (Transient)               (Transient)            (Washout)

This hierarchical structure yields non-intuitive clinical insights. For instance, Pacheco et al. demonstrated through compartmental modeling that targeted therapies can successfully cure chronic myeloid leukemia by eliminating downstream proliferating cells, even if the therapy never successfully kills a single underlying leukemic stem cell.

Similarly, compartments can represent distinct anatomical environments. To predict patient survival in metastatic breast cancer, Ascolani et al. constructed a multi-compartment system tracking the physical journey of circulating tumor cells as they escape the primary tumor, enter the bloodstream, and ultimately colonize the bone.

Capturing Time Lags and Feedback Loops

In nature, biological processes are not instantaneous. It takes time for a cell to mature, for a virus to replicate inside a host, or for a hormone signal to trigger a response. When these temporal lags interact with negative feedback loops, they can cause the entire system to destabilize and oscillate.

To capture these dynamics, the modeling framework expands from simple ODEs to delay-differential equations (DDEs) and integro-differential equations:

  • Maturation Lags: Drobnjak developed a system of delay-partial differential equations that explicitly accounts for cell age and maturation status in a single cell lineage, identifying three distinct biological mechanisms that trigger wild oscillations in blood cell counts.
  • Feedback Delays: In modeling platelet production, Langlois et al. used integro-differential equations with constant delays to separate development into discrete phases like mitosis and endomitosis. This mathematical formulation allowed them to demonstrate how dysfunctions in hormone receptors alter the feedback loop, explaining the cyclical nature of certain blood disorders.

By embedding time delays directly into the equations, these models capture the momentum inherent in biological systems, explaining behaviors that look chaotic but are actually tightly regulated feedback loops running on a delay.

Simulating the Race of Infection and Evolution

In virology, compartment models map out the high-stakes race between viral replication and host immunity. These models can operate at multiple scales simultaneously. For example, Mohammadi et al. used ODEs to partition the intracellular HIV lifecycle into distinct phases based on the accumulation of replication intermediates, while Wang 0001 et al. integrated intracellular RNA replication with intercellular spread to explain why COVID-19 patients experience such highly diverse clinical outcomes.

Because these models are rooted in physical principles like conservation of mass and population kinetics, they can make long-term projections from short-term clinical data. Alexandre et al. leveraged ODEs of antibody production and decay to project the 15-year durability of an Ebola vaccine using data gathered from brief clinical trials.

Finally, compartment models excel at simulating evolutionary pressures in real time. By setting up competing compartments for different viral genotypes, researchers can observe how interventions shift the fitness landscape. Savković et al. used this method to track how gene therapies pressure viruses to switch strains, while Gopalakrishnan et al. integrated statistical mutational pathways with mechanistic viral dynamics to predict how the delicate trade-off between drug resistance and replication efficiency ultimately dictates whether a treatment succeeds or fails.

Key Foundational Work

On Modeling HIV and T Cells In Vivo: Assessing Causal Estimators in Vaccine Trials (2006) — contributed 1 concept elements to this direction.
Oscillations in a Maturation Model of Blood Cell Production (2006) — contributed 1 concept elements to this direction.
Distributed Object Replication Strategies for a Remote Caching Architecture (1992) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_mechanistic_viral_dynamics_modeling_48bc47c6,
  title = {mechanistic viral dynamics modeling},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/mechanistic-viral-dynamics-modeling/rdr_48bc47c6929b466b8e4211043d59a319},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_mechanistic_viral_dynamics_modeling_48bc47c6_v1,
  title = {mechanistic viral dynamics modeling},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/mechanistic-viral-dynamics-modeling/rdr_48bc47c6929b466b8e4211043d59a319/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}

See something wrong? Report a mistake