Mantis Biotech Uses Digital Twins to Revolutionize Biomedical Research

Large language models are becoming increasingly important in the field of biomedical research. Trained on vast datasets, these models offer the potential to accelerate key areas such as genomics, clinical documentation, diagnostics, and drug discovery. However, the lack of comprehensive datasets, particularly for rare diseases and unusual conditions, often impedes progress.

Mantis Biotech, a New York-based company, presents a novel solution to this challenge. The company has developed a platform that combines diverse data sources to generate synthetic datasets, ultimately creating "digital twins"—physics-based, predictive models of human anatomy, physiology, and behavior.

• These digital twins can be utilized for:
  • Data aggregation and analysis.
  • Studying new medical procedures.
  • Training surgical robots.
  • Simulating and predicting medical and behavioral issues.
• Georgia Witchel, founder and CEO, highlighted the potential of this technology in sports, particularly for predicting injuries in players.

Mantis' sophisticated platform draws information from textbooks, cameras, sensors, and medical imaging, employing an LLM-based system to process this data. The platform's physics engine is particularly crucial, providing realistic modeling of anatomy by correctly mimicking physical attributes.

• It solves problems of scarcity in datasets by generating required data easily, like creating hand-pose data for amputees.
• Applications include improving access to biomedical data for better training AI models without violating privacy norms.

Mantis has seen success, especially in professional sports teams like the NBA, and has recently secured $7.4 million in seed funding from Decibel VC, Y Combinator, and other investors. The funds will assist in marketing and expansion efforts, eventually aiming to offer the platform for preventive healthcare and FDA trials.

The vision: By leveraging digital twins, Mantis aspires to broaden the scope of medical research and predictive modeling in health sectors, offering insights while maintaining ethical use of data. The company plans to broaden its outreach to pharmaceutical labs and public healthcare sectors.