Recursion Pharmaceuticals

Recursion is advancing the first potential drug for rare disease FAP, discovered with AI.

For FAP Awareness Week, we’re resharing our video featuring CEO and President Najat Khan announcing the first clinical validation of Recursion’s AI-enabled drug discovery platform – positive Phase 1b/2 data for AI-identified drug REC-4881 in Familial Adenomatous Polyposis (FAP).

This progressive rare disease – which often begins in the teens and results in hundreds or thousands of polyps developing in the GI tract which have a 100% likelihood of becoming cancerous if not removed, requires a lifetime of invasive surveillance and life-altering surgeries. There are currently no FDA-approved treatments.

Recursion used AI not only to deliver the first potential treatment for this disease but also to quantify – for the first time at scale – the true natural progression of polyp burden in the trial-relevant population. This program is the first clinical validation of the Recursion OS, and demonstrates our ability to translate unbiased phenotypic insights into potentially differentiated treatments for diseases with high unmet need.

We’re working hard to advance new solutions for FAP patients, this week and every week.

🧬 Bridging the gap in AI drug discovery.

Ali Denton, Staff Machine Learning Scientist at Recursion and one of the authors on the recent paper in Nature Biotechnology, explains how the AI model TxPert predicts how a cell will respond to perturbations.

Predicting a cell’s RNA activity, or transcriptome, is key to bridging the gap between cellular changes and clinical outcomes and advancing the potential for AI drug discovery. As Ali says, “with hundreds of cell types and so much disease variation, the total possibilities are too vast to measure in a lab.”

She describes how TxPert allows us to perform a “Virtual Assay,” taking the mathematical signature of a healthy cell called the Basal State and adding the perturbation’s embedding to deliver a highly accurate prediction of what the cell’s transcriptome will look like after treatment.

TxPert uses layered graph-based models that integrate phenomics — or how a cell looks — and transcriptomics — which genes are expressed — along with massive public biological knowledge resources.

The model can even predict how a perturbation will work in entirely new cell lines it hasn’t seen before as well as accurately forecast the effects of “double perturbations,” consistently identifying “unknown unknowns” that traditional models — and even massive general-purpose AI — often miss.
Ali notes that TxPert is currently predicting genetic perturbations, but more flexible models — including those predicting drug effects — are in the works.

👉 Check out the full paper in Nature Biotech: https://www.nature.comarticles/s41587-026-03113-4