Nat. Biotechnol.｜生物技术领域的数字孪生困境

TL;DR - A Nature Biotechnology feature surveys digital twins in biotechnology, highlighting their potential in drug discovery and clinical trials alongside unresolved definition and validation problems. Narrow, purpose-built models currently appear more practical than complete virtual replicas of patients.

• Three approaches dominate: mechanism-based causal models, data-driven clinical-trial simulations, and organ-chip systems coupled with pharmacological modeling.
• Applications include identifying drug targets, stratifying patients, predicting control-group outcomes, and reducing clinical-trial enrollment requirements.
• Biological uncertainty, incomplete mechanistic knowledge, black-box predictions, and inconsistent definitions complicate validation and trust.
• Progress depends on use-specific evaluation standards, uncertainty reporting, and transparency about each model’s limitations.