Biotech Startup Tahoe Therapeutics Raised $30 Million To Build AI Models Of Living Cells
Tahoe Therapeutics, a biotech startup co-founded by Kevin Shokat, Nima Alidoust, Johnny Yu, and Hani Goodarzi, is making significant strides in the field of biology by working towards digitally simulating a living cell. The ultimate goal is to use computers to accurately predict how new medicines will interact with the body, thus improving the testing process on animals and humans.
While advancements in large language models have helped in modeling protein behavior, simulating the complexities of an entire cell has proven to be more challenging due to a lack of sufficient data. However, Tahoe Therapeutics made a breakthrough earlier this year with the release of Tahoe-100M, a dataset containing 100 million datapoints showcasing how different cancer cells responded to interactions with over 1,000 molecules. This data, known as perturbations, is crucial for training AI models to predict how cells will react to various substances.
The company’s Mosaic platform allows them to combine cells from diverse patients and organs, generating extensive single-cell atlases that provide valuable insights into how drugs affect different individuals. This approach sets Tahoe apart from conventional techniques that test cells from one person at a time.
Tahoe’s ability to produce massive datasets quickly is their key strength, enabling them to build virtual cell models for drug discovery. Recently, the company secured $30 million in new funding, led by Amplify Partners, bringing their total funding to $42 million and valuation to $120 million.
Krish Ramadurai, a partner at AIX Ventures and a board member at Tahoe, emphasized the importance of accurate AI predictions in biotech and commended Tahoe for their data generation capabilities, which enhance the precision of new models.
Following the publication of the Tahoe-100M dataset, the non-profit organization Arc Institute developed an open-source virtual cell model called State, incorporating Tahoe’s data for training. State demonstrated twice the accuracy of other AI models and outperformed simpler machine learning programs.
Tahoe’s journey began nearly a decade ago when cofounder Johnny Yu developed the technology for Mosaic while working at the University of California San Francisco. Together with Nima Alidoust, Hani Goodarzi, and Kevin Shokat, they founded Tahoe to pursue the vision of building virtual cell models for drug discovery.
With their latest funding, Tahoe aims to expand their dataset to over a billion single-cell datapoints to accelerate the development of new cancer treatments. While they plan to keep their larger datasets proprietary, they are considering collaborations with pharmaceutical or AI companies to leverage their data for developing new medicines or AI models.
In conclusion, Tahoe Therapeutics is at the forefront of revolutionizing biology with its innovative approach to generating extensive datasets for virtual cell modeling. As they continue to advance their technology and build partnerships, the company remains committed to driving progress in the field of drug discovery.
