Recursion and Enamine Release AI-Curated Screening Libraries Targeting 100 Challenging Drug Targets

Recursion and Enamine announced the release of ten curated screening libraries generated through their AI-enabled collaboration. The libraries are designed to accelerate hit identification across 100 difficult-to-target areas of human biology and represent a filtered subset of more than 15,000 newly synthesized compounds.

The effort combines Recursion’s proprietary machine learning platform—the Recursion OS—with Enamine’s REAL Space, a virtual chemical collection of nearly 65 billion make-on-demand small molecules.

The effort focused on narrowing down this large chemical space to a manageable set of compounds predicted to interact with 100 preselected and biologically relevant targets. The selection process used Recursion’s predictive models to identify candidate molecules expected to have functional activity against targets not currently addressed by its internal 1.2 million-compound library. The resulting compound sets are smaller and more focused than traditional high-throughput screening (HTS) collections, and are intended to increase efficiency in early drug discovery campaigns.

Recursion's NVIDIA-powered AI supercomputer. Source: Recursion

The project also provided Recursion access to tens of thousands of compounds from the REAL database for internal research. In exchange, Recursion contributed to the development of compound libraries that Enamine is now preparing to offer more broadly for hit-finding applications.

While Enamine's REAL platform has previously been used in virtual screening campaigns, this marks its first application at this scale across a large panel of targets, leveraging AI to guide compound selection. According to the release, compounds from the curated sets can be synthesized within 3–4 weeks with high feasibility and at low cost, due to Enamine’s parallel synthesis framework and 143,000 building blocks.

Topics: Manufacturing & Pharma 4.0