Otter Biosciences trains artificial intelligence models on living human skin tissue

Otter Biosciences is working to reshape biological compound discovery by training artificial intelligence models directly on living human skin tissue obtained from plastic surgery procedures, seeking to bypass the slow pace of traditional laboratory experiments and the inaccuracies of animal models and synthetic alternatives in mimicking real human organs.

Founded in 2022 by Michael Polansky alongside Kyung-Jin Jang, Chris Hinojosa, and Stanley King, the startup set out to bridge the gap between rapid software innovation and the slow pace of biological and chemical research, where ethical constraints prevent initial testing directly on humans and common laboratory alternatives like simplified cell cultures fail to reflect human biology with sufficient precision.

The company relies on a proprietary biological support system that supplies living human tissue with essential nutrients and clears metabolic waste, keeping it viable outside the body for up to a full month compared to standard laboratory timeframes of only a few days.This extended window allows researchers to monitor slow biological processes such as pigmentation changes, collagen regeneration, and skin barrier repair, as well as track stress and inflammatory responses and gradual recovery pathways after exposing samples to ultraviolet radiation that simulates sunburn.

The company procures discarded surgical skin tissue through certified biobanks and brokers operating under institutional review board oversight with documented donor consent, fully de-identifying data and covering retrieval costs rather than purchasing tissue outright. The entire process runs in a closed loop, where the AI model predicts promising chemical compounds, those predictions are tested on living tissue, and the experimental data is fed back into the model to refine subsequent predictions.

This approach shortened the discovery cycle from manual literature screening, which took 18 months to yield two leads, to generating a new chemical candidate roughly every six weeks, with six active programs currently in the pipeline. That pace stands out in an active ingredients market where only about 200 research-backed compounds are currently approved.

The startup is currently focused on discovering cosmetic ingredients that meet internationally recognized OECD guidelines and licensing them to commercial brands, alongside research partnerships that include studying skin rashes caused by oncology drugs.The company has raised about 23 million dollars from investors including Wing Ventures, Initialized, and Hawke Tale, and runs its models on dedicated local servers to protect the privacy of its biological data away from the cloud.