Tahoe Therapeutics

EXCLUSIVE: Tahoe Therapeutics, a 15-employee South San Francisco biotech, has developed the largest virtual cell model to date. And it's giving it away for free. I talked with CEO Nima Alidoust, CSO Johnny Yu, and ML director Shreshth Gandhi about why they plan to open-source Tahoe-x1 tomorrow, and what's next for virtual cell models and for Tahoe. The biotech raised a $30 million Series A earlier this year at a valuation and already is eyeing the clinic. Alidoust told me they are "getting very close to putting together an IND package for our lead program." More here in my latest exclusive for Endpoints News: https://lnkd.in/emDrAaNg

We are excited to share Tahoe-x1 with the world tomorrow! And thanks Andrew Dunn for the coverage.

We at Tahoe Therapeutics believe models that live up to the title "virtual cell" are a few years away. To build them, we need data. And we need open source, compute-efficient models that make the search for the ultimate models an empirical problem. Our Tahoe-100M dataset was a step towards solving the data challenge. Tomorrow, we will open source Tahoe-x1, a 3B parameter single-cell foundation model, as a step towards solving the modeling challenge. In anticipation of that, we had a chat with Andrew Dunn from Endpoints News. Enjoy!

🧬 Tahoe Workshop at scverse conference 2025 🧬 Disease Biology and Therapeutics in the Age of Frontier Datasets How are large-scale perturbation datasets transforming research in disease biology and drug discovery? This session will focus on how large-scale perturbation datasets are enabling new capabilities in disease biology and translational science — moving from static maps to predictive models and controllable systems. 📅 November 17-19, 2025 at Stanford University, California 🏛️ Format: 2 days of main conference + 1 day of satellite workshops across the Bay Area ℹ️ More info: https://lnkd.in/dPc6sMHf 🚀 Ready to join us? Register: https://lnkd.in/dTRydMwj Tahoe Therapeutics #scverse2025 #ComputationalBiology #DrugDiscovery #TahoeTherapeutics #SingleCell #PerturbationBiology

🌟 Come see what we’re building at Tahoe Therapeutics at #ASHG25! 🌟 We’re proud to share that Airol Anne U., co-first author on our Tahoe-100M paper and a true builder 🔨 on the Mosaic platform, will be presenting our latest work — including unpublished data that pushes the boundaries of single-cell discovery. Airol has been instrumental in making this project a reality. Her creativity, technical rigor, and perseverance embody what makes Tahoe so special. 💪 📍 Today at 2:30 PM — Board 4101F Come by to meet Airol, learn more about our next generation of large-scale single-cell datasets, and see what’s coming next from Tahoe Therapeutics. 🧬 #ASHG25 #singlecell #genomics #Mosaic #Tahoe100M #TeamTahoe

What if we combined genetic & chemical perturbations to infer previously unknown mechanisms of action for any given compound? Compound perturbations in Tahoe-100M aligned with CRISPR data allowed us to do that at scale. In our new blog post, we share an example of how powerful this combination can be. We paired Tahoe-100M with Myllia’s CROP‑seq: 218 CRISPR targets in A549. We unify drug & gene maps via gene‑set scoring (MSigDB, Vision), enabling direct comparison of drug‑ vs gene‑induced signatures. The paradigm shift: treat drugs as programmable phenotypic levers—design compounds to recreate desired transcriptional states. This blueprint—uniting CRISPR genetics with small‑molecule perturbations—fuels our upcoming 1B+ cell atlas: more cells, more contexts, greater confidence in mechanism & phenocopy. Read more in the full post about how we used it to recapitulate target pathways of compounds with known MoA, uncovered new MoAs for other compounds, and how the method generally performs. Link in comments.

Join us on September 4th at MBC BioLabs San Francisco for our next Founders at Work Speaker Series: Tahoe Therapeutics on “Building Frontier Datasets to Power the Virtual Cell,” featuring Johnny Yu, CSO & Co-founder. Generation of large scale single-cell datasets are now integral to the vision of the Virtual cell. Tahoe Therapeutics is leading a generation of frontier scale datasets across diverse perturbations to i) develop programs from unbiased datasets ii) enable modeling of the cell states. Please register here to save your spot ➡️ https://lu.ma/wkx2xh3g

We are thrilled to have the chance to support Nima Alidoust, Johnny Yu, Hani Goodarzi, Kevan Shokat and the entire Tahoe Therapeutics team as a part of their $30M Series A financing. We firmly believe that by advancing deep biological insights, Tahoe Therapeutics can significantly expedite the discovery and development of meaningful medicines for people affected by cancer and other severe conditions. With this financing, Tahoe aims to generate an 1 billion single-cell datapoints, paving the way for unprecedented drug-patient mapping at a scale, and to selectively partner in enabling access to their foundational model for virtual cells. We are filled with immense optimism about the profound impact that this work will have on the medical field.

Ask any scientist about curing cancer and you'll get mixed signals. Good news: cancer death rates have dropped dramatically over 20 years. Bad news: 95% of cancer drugs still fail in clinical trials. The translation problem is brutal: a drug works on a single protein → fails in a cell → fails in mice → fails in humans. We’re designing more molecules than ever, but only 1 in 10,000 that look promising actually help patients. Basically, we're great at designing molecules that fail spectacularly in humans. Most AI biotech companies focus on the top of the funnel - generating more drug candidates faster. But if 95% of these "successes" will still fail, the bottleneck isn't creating more candidates; it's predicting which ones actually work in patients. Tahoe Therapeutics is solving the 95% Problem by turning cancer into a data problem. Instead of testing molecules one by one in isolation, they're building comprehensive digital models of how drugs interact with real human biology. The concept: Virtual Cells. Digital representations of life's fundamental unit that can predict drug responses without manual lab testing. But building these models requires massive amounts of high-quality data showing drug-cell interactions across diverse patients. Tahoe recently released the game-changing Tahoe 100-M dataset: 100M drug-cell interactions and 60,000 drug-patient interactions. It's 50x larger than all public drug-perturbed data combined. LLMs have the entire internet corpus - now scientists have Tahoe-100M. How did they build it? Two breakthrough innovations: "Mosaic tumors" - pooling cells from diverse cancer patients into composite tumors that represent broad genetic spectrums "Growth balancing" - mixing cells in proportions that prevent any one type from dominating Then there’s the tracking challenge: how do you measure each cell's reaction in this multi-cell mix? Most labs use artificial barcoding, but Tahoe discovered natural genetic barcodes - unique patterns that let AI map every cell back to its original patient without modification. The results are already promising. Using models trained on Tahoe-100M, they've identified drug combinations targeting 40% of colorectal cancer patients (vs typical 10-15%). What's fascinating: their discoveries contradicted veteran researchers' intuitions about patient selection. These ML models aren't just finding needles in haystacks - they're revealing we've been looking in the wrong haystacks entirely. Human intuition shaped by limited datasets might be systematically missing patterns visible only at sufficient scale. Yesterday Tahoe announced a $30M Series A led by our team at Amplify Partners. Their next milestone: 1 billion cellular data points by the end of 2026. We're thrilled to support Nima Alidoust, Johnny Yu, Hani Goodarzi, and Kevan Shokat as they chart a new course for precision medicine. Link to full post in comments.