Virtual Organisms and Digital Biology
Virtual-organism claims are the natural next step after virtual-cell claims, and the most dangerous place to overstate what current systems support. The gap between a cell-state predictor and an organism-scale simulator includes development, tissue interactions, physiology, immune function, behavior, environment, and time.
- Distinguish virtual cells, organoids, digital twins, and virtual organisms
- Identify what measurements are required for organism-scale simulation
- Recognize when phenotype prediction exceeds available evidence
- Evaluate whether a simulator has been validated outside calibration data
- Keep mechanistic and statistical models separate when interpreting claims
Introduction
The phrase “digital twin” is easy to apply and hard to earn. In biology, a useful digital twin must specify the represented system, the input data, the modeled mechanisms, the update cycle, the uncertainty, and the decision it supports. A cell-state model, organoid model, animal model, and patient-specific clinical model all deserve different evidence standards.
Whole-cell modeling provides a serious precedent in a minimal organism (Karr et al., 2012). Virtual-cell roadmaps and benchmarks show the current cell-scale ambition (Bunne et al., 2024; Roohani et al., 2025). Moving from those anchors to organisms multiplies the validation burden.
Demonstrated
Demonstrated capability includes narrow simulations of defined cellular or organismal subsystems, especially when mechanisms and parameters are curated and validated against measured phenotypes.
Demonstrated capability also includes digital models used for experimental planning when the decision is local: choosing a perturbation, prioritizing an assay, or testing whether a mechanism is internally consistent.
Theoretical
Theoretical capability includes organism-scale models that combine mechanistic simulation with foundation-model representations. This direction is plausible, but credibility depends on measured transfer across tissues, developmental states, and environmental conditions.
Theoretical capability also includes organoid and model-organism simulations that support preclinical triage. The evidence standard should include prospective experiments, not only retrospective fit.
Beyond current capabilities
Beyond current capabilities includes a general virtual organism that forecasts phenotype from genome and environment with enough reliability to replace animal, organoid, or field experiments.
Beyond current capabilities also includes patient-specific biological twins for treatment decisions without clinical validation. That question belongs partly in clinical AI, but the biological-modeling burden begins here.
Practice Notes
- Define the modeled system before judging the model.
- Separate whole-cell, organoid, animal, and patient-specific claims.
- Require prospective validation for any intervention-ranking use.
- Report uncertainty at every scale transition.
- Treat replacement-of-experiment claims as unsupported unless replacement has been tested.