Mantis Biotech is making ‘digital twins’ of humans to help solve medicine’s data availability problem
Bridging Data Gaps in Medicine with Digital Twins
Large language models have revolutionized biomedical research, clinical documentation, diagnostics, and drug discovery. However, these models often struggle when reliable, representative data is limited—particularly in cases like rare diseases or unusual health conditions.
Mantis Biotech’s Solution: Synthetic Human Data
New York-based Mantis Biotech aims to tackle this issue by generating digital twins of humans. Their platform aggregates data from various sources, such as textbooks, biometric sensors, motion capture, training logs, and medical imaging. Using an advanced language model, the platform validates and synthesizes this data, which is then processed by a physics engine to create detailed, predictive simulations of human anatomy, physiology, and behavior.
Applications and Impact
Mantis's digital twins can be used for a variety of applications, including:
- Testing and studying new medical procedures
- Training surgical robots
- Simulating and predicting injuries or medical events—like forecasting an athlete’s risk for specific injuries based on their training and health data
The ability to generate synthetic datasets—for example, modeling a hand with missing fingers—enables research and innovation even when real-world data is scarce or protected for privacy reasons.
Success in Sports and Beyond
Mantis Biotech has already found success working with professional sports teams, providing digital twins to monitor athlete performance and predict health outcomes. The company recently raised $7.4 million in seed funding, led by Decibel VC and supported by investors including Y Combinator.
Looking Ahead
As Mantis Biotech continues to develop its technology, it aims to serve not just athletes but also healthcare providers, pharmaceutical labs, and researchers. The ultimate goal is to make these digital twins accessible for preventative healthcare and clinical research, helping unlock insights where real data is hard to obtain.
For more details, read the original article by Ram Iyer at TechCrunch.
