Mantis Biotech Creates Digital Twins to Solve Medical Data Gaps

The potential for large language models to revolutionize biomedical fields is immense, from accelerating drug discovery to enhancing clinical diagnostics. However, a significant barrier remains: these AI systems often falter when faced with rare diseases or unusual medical conditions where comprehensive, reliable data simply doesn’t exist. A New York startup, Mantis Biotech, is tackling this fundamental data availability gap by constructing highly detailed digital twins of the human body.

Mantis’s approach involves integrating a wide array of disparate data sources, including medical textbooks, biometric sensors, motion capture feeds, and training logs. Its platform employs an LLM-based system to route, validate, and synthesize these streams. The crucial next step is processing everything through a proprietary physics engine, which creates high-fidelity, predictive models of human anatomy, physiology, and behavior. This process generates the synthetic datasets needed to build accurate virtual representations.

CEO Georgia Witchel emphasizes that the physics layer is essential. It grounds the synthetic data in realistic biomechanics, allowing the platform to model scenarios for which real-world data is absent. “If I asked you to do hand-pose estimation for someone who is missing a finger, it would be really, really hard, because there are no publicly available datasets of labeled hand positions of someone who is missing a finger,” Witchel explained. “We could generate that dataset really, really easily, because we just take our physics model and we say, remove finger X, regenerate model.”

These predictive models are designed for diverse applications, such as studying new medical procedures, training surgical robotics, or simulating potential health issues. In one described use case, a sports organization could forecast an NFL player’s risk for an Achilles injury by analyzing a digital twin that incorporates their performance metrics, training load, diet, and activity history. The technology is particularly suited for edge cases in healthcare, where ethical and regulatory constraints often limit access to patient data for AI training.

Witchel envisions a paradigm shift in how researchers interact with human data. “You know how when you see a three-year-old running around, and they have a Barbie, and they’re holding it by one leg and smashing it against a table? I want people to have that mindset with our digital twins,” she said. This approach, she argues, enables rigorous testing and exploration using virtual humans while fully respecting individual privacy and avoiding the exploitation of personal health data.

Currently, Mantis has found a strong early market in professional sports, where modeling elite athletes is a priority. The startup counts an NBA team among its main clients, creating digital representations that track an athlete’s biomechanics over time and correlate them with variables like sleep and specific movements. To fuel its expansion, Mantis recently secured $7.4 million in seed funding led by Decibel VC, with participation from Y Combinator and other investors.

The capital will support hiring and go-to-market efforts. Looking ahead, the company plans to further develop its core technology with the goal of releasing a public platform focused on preventative healthcare. Mantis is also engaging with pharmaceutical labs and researchers conducting FDA trials, aiming to provide deeper insights into patient treatment responses through advanced simulation.