Tahoe-100M for Cancer Research with Aura
Making large-scale functional genomics accessible to cancer researchers through natural language and AI-powered insights.
The pace of biomedical discovery is limited not by data but by the ability to navigate it. At Fastfold, our mission is to build AI agents that democratize access to complex scientific data and accelerate therapeutic research. Today, we’re excited to announce support for the Tahoe-100M dataset in Aura, our AI-powered scientific agent.
Tahoe-100M is a giga-scale single-cell perturbation atlas developed by Tahoe Therapeutics (formerly Vevo) in partnership with Arc Institute. It consists of over 100 million single-cell transcriptomes, sampled from 50 pooled human cancer cell lines treated across 1,200 unique small-molecule perturbations .
This dataset captures a multi-modal snapshot of cancer biology, including:
- Transcriptomic profiles across perturbation conditions,
- Drug-cell response data spanning 60,000+ drug–dose–cell line interactions. Arc Institute Release
- Rich metadata on driver mutations, tissue origin, and treatment specifics.
Together, this forms the most comprehensive publicly available functional atlas at single-cell resolution, enabling mechanistic insight and AI-driven modeling at an unprecedented scale.
Using Tahoe with Natural Language
With Aura, scientists can explore Tahoe-100M using plain English. For example:
- “Help me find top genes for Skin with Loss of Function drivers.”
- “Compare TP53 gene expression levels in Skin-derived cell lines with and without LoF drivers, using DMSO_TF as the control.”
Aura returns curated subsets, visualizations, and statistical comparisons, reducing days of computational setup to seconds of inquiry.
Why It Matters
As cancer research becomes increasingly data-intensive, accessibility remains a major challenge. Tahoe-100M offers vast complexity—but its richness is hard to unlock without specialized tools. Aura bridges this gap.
By combining Tahoe’s scale with natural language querying, Aura empowers researchers—regardless of coding expertise—to run custom, reproducible, and interpretable analyses in seconds.
Our Commitment
This integration reflects our broader vision: to accelerate bioscience through intuitive, AI-native tools that are powerful and scientifically rigorous. Whether investigating gene-drug interactions, mutation landscapes, or expression patterns, Aura is designed to assist transparently and reproducibly.
Big thanks to the teams at Tahoe Therapeutics and Arc Institute for building and sharing this incredible dataset. Their work is raising the bar for what’s possible in large-scale functional genomics
Interested in using Aura with Tahoe-100M in your research or lab? Get in touch with us.
References
Tahoe-100M: A Giga-Scale Single-Cell Perturbation Atlas for Context-Dependent Gene Function and Cellular Modeling
Jesse Zhang, Airol A. Ubas, Richard de Borja, Valentine Svensson, Nicole Thomas, Neha Thakar, Ian Lai, Aidan Winters, Umair Khan, Matthew G. Jones, Vuong Tran, Joseph Pangallo, Efthymia Papalexi, Ajay Sapre, Hoai Nguyen, Oliver Sanderson, Maria Nigos, Olivia Kaplan, Sarah Schroeder, Bryan Hariadi, Simone Marrujo, Crina Curca, Alec Salvino, Guillermo Gallareta Olivares, Ryan Koehler, Gary Geiss, Alexander Rosenberg, Charles Roco, Daniele Merico, Nima Alidoust, Hani Goodarzi, Johnny X. Yu
bioRxiv, posted Feb 20, 2025. DOI: 10.1101/2025.02.20.639398