Recursion leverages proprietary data—including extensive cell microscopy data, transcriptomic, and patient data, and an AI-enabled platform to accelerate the discovery and development of transformative new medicines for dozens of diseases with unmet medical need. The company, headquartered in Salt Lake City, Utah, with additional sites in New York City, London, the Oxford area, and Montreal, has built a drug discovery platform that integrates automated chemistry, automated biology, and cloud computing tools to uncover novel therapeutic targets and precision-design new drugs, potentially cutting the time to discover and develop a new medicine by a factor of 10.

It's a paradox which observers call "Eroom's Law": drug discovery, despite improvements in medical technology, has become increasingly slower and more expensive. As a result, the number of FDA-approved new medicines reaching patients has stagnated, while the costs of pharmaceutical R&D has continued to increase.

There are approximately 6,000 rare diseases affecting an estimated 25 million people—many of them young children—in the United States alone. The high cost and extremely long timelines of conventional drug development are even more daunting for rare diseases. As a result, less than 5 percent of rare diseases have FDA-approved treatments.

Meanwhile, 1 in 5 people will develop cancer in their lifetime, and approximately 1 in 5 cancer diagnoses in the US is considered a rare cancer. 

Recursion is positioned to rapidly improve the development and success rate of new treatments by combining a data-driven understanding of biology and chemistry with the latest in machine learning tools to reimagine the drug discovery and development process end to end—from target discovery informed by patient genetic data to clinical trials optimized to include the patient populations most likely to benefit. The company is advancing a pipeline of clinical stage drugs—and its lead drug, REC-4881, is a potential first-in-disease treatment for the rare disease Familial Adenomatous Polyposis (FAP) that has shown positive Phase 2 results and is one of the most advanced AI drugs in development today.

In its quest to transform how medicines are made, engineering, AI-pattern recognition, and cloud computing have proven as necessary as biochemical expertise. Recursion's data pipeline incorporates image processing, inference engines, and deep learning modules. Its engineers have built a platform that supports bursts of computational power that weigh in at trillions of calculations per second.

Recursion has integrated a massive processing pipeline and neural networks into a target platform that is scalable, cost-effective, and tracking to achieve potential treatments for both rare and common conditions in cardiology, neurology, dermatology, oncology, immunology, and ophthalmology, among others.

It starts with wet biology—plates of glass bottom wells containing thousands of healthy and diseased human cells. The firm's biologists run experiments on the cells, applying stains that help characterize and quantify the features of the cellular samples: their roundness, the thickness of their membrane, the shape of their mitochondria, and other characteristics.

A microscopy team captures this data by snapping high-resolution photos of the cells at several different light wavelengths. Google Cloud allows this massive store of data to be readily available when needed for Recursion's machine learning models. A data pipeline that sits on top of Google Kubernetes Engine (GKE) running on Google Cloud, extracts and analyzes cell features from the images. Mathematical models with data that represent the cell features are deployed in packages to GKE containers. Then, data is processed by deep neural networks to find patterns, including those humans might not recognize. The neural nets are trained to compare healthy and diseased tissue signatures with those of tissues before and after a variety of drug treatments. This strategy returns new approaches for treating diseases—based entirely on data—that were previously unknown, as well as the best way to design molecules to be effective drugs, and the most promising patient populations (and where to find them).

In over a decade since its founding, Recursion has expanded and improved its end-to-end operations, including automated high throughput and automated chemistry labs, and built one of the largest proprietary biological and chemical datasets—over 50 petabytes—including microscopy images (phenomics) and RNA changes (transcriptomics), integrating this with patient genetic and health data. "We generate massive amounts of fit-for-purpose data and then use modern AI tools to make sense of that complexity, building powerful maps of biology that turn discovery into a computationally searchable problem," says Hayley Donnella, PhD, VP of Frontier Research at Recursion. "The scale of the data is significant, but what matters more is that the data is complementary and connected." 

As the company has scaled, they've had to rethink their approach to managing big data. Google Cloud has been an important partner. "I think of Google Cloud as the foundation, the walls, and then within those, we're able to build our own apartments that are our models or our tools," says Maureen Makes, VP of Engineering at Recursion. "We also leverage TPUs for our model inference to quickly and reliably have outputs from this data on demand. We've seen about a 50% reduction in price while keeping the same quality."

The company has also leveraged Google Cloud to release open source foundation models like Open Phenom-S/16 for life sciences companies to assess the effectiveness of drugs in development using Recursion's microscopy data, setting a new gold standard for the industry.