Large language models (LLMs), trained on extensive datasets, have the potential to significantly enhance genomics research, streamline clinical documentation, improve real-time diagnostics, support medical decision-making, accelerate drug discovery, and even generate synthetic data for scientific experiments.
However, the promise of these models to revolutionize biomedical research often faces a hurdle: their effectiveness diminishes in less common scenarios, such as rare diseases or unusual conditions, where sufficient and representative structured healthcare data is scarce.
New York-based Mantis Biotech claims it is developing a solution to address this data scarcity. The company’s platform integrates various data sources to create synthetic datasets, which are then used to build “digital twins” of the human body. These digital twins are physics-based, predictive models of anatomy, physiology, and behavior.
Mantis is promoting these digital twins for comprehensive data aggregation and analysis. They could be applied to study and test new medical procedures, train surgical robots, and simulate and forecast medical issues or even behavioral patterns. For instance, as Mantis founder and CEO Georgia Witchel explained to TechCrunch, a sports team could predict the likelihood of an NFL player developing an Achilles heel injury based on their recent performance, training intensity, diet, and career longevity.
To construct these twins, Mantis’s platform first gathers data from diverse sources like textbooks, motion capture cameras, biometric sensors, training logs, and medical imaging. Subsequently, an LLM-based system processes this information to route, validate, and synthesize the data streams. This combined data is then run through a physics engine to generate high-fidelity representations of the dataset, which can then be used to train predictive models.
Witchel stated, “We’re able to take all these disparate data sources and then turn them into predictive models for how people are going to perform. So anytime you want to predict how a human being is going to be performing, that is a really good use case for our technology.”
Witchel emphasized to TechCrunch that the physics engine layer is crucial because it enhances the available information by anchoring the generated synthetic data and realistically simulating the physics of human anatomy.
She gave an example: “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. We could generate that dataset really, really easily, because we just take our physics model and we say, remove finger X, regenerate model.”
Given Mantis’s ability to fill gaps in data sources, Witchel believes the platform has broad applicability across the biomedical industry, where information on procedures or patients is often hard to access, unstructured, or siloed. She specifically highlighted its value for edge cases or rare diseases, where data is scarce due to ethical and regulatory restrictions on using patient data in public datasets or for training AI models.
“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 expressed. “I think that’s going to open up people to this idea that humans can be tested on when you’re using virtual humans. I feel currently, people operate with the exact opposite mindset, which totally makes sense, because people’s privacy should be respected. In fact, I don’t really think people’s data should be exploited at all, especially when you have these digital twins.”
Currently, Mantis has found success in professional sports, likely due to the inherent need to model high-performing athletes. Witchel mentioned that an NBA team is among the startup’s primary clients.
She elaborated, “We create these digital representations of the athletes, where it basically shows here’s how this athlete has jumped, not just today, but for every single day in the past year, and here’s how their jumps are changing over time compared to the amount that they’re sleeping, or compared to how many times they lift their arms above their head.”
The startup recently secured $7.4 million in seed funding, led by Decibel VC, with contributions from Y Combinator, several angel investors, and Liquid 2. This capital will be allocated towards hiring, advertising, marketing, and go-to-market initiatives.
Witchel indicated that Mantis’s next steps include further developing its technology and eventually launching the platform to the public, focusing on preventative healthcare. The company is also working to serve pharmaceutical labs and researchers conducting FDA trials, aiming to provide insights into patient responses to treatments.