AI-predicted structures reveal 'superdark' protein resembling GPCR but functioning differently

September 9, 2026 at 05:00 AM1 min

This digest was compiled by AI from multiple sources — links to the originals are below.

Researchers mining millions of AI-predicted protein structures identified an evolutionarily conserved protein that resembles a G-protein-coupled receptor but shows unconventional behavior. The finding, reported in Nature on September 9, 2026, sheds light on the 'dark proteome' of proteins with unknown functions.

Key Facts

Lee et al. reported the discovery in Nature on September 9, 2026.
The protein is evolutionarily conserved and resembles a G-protein-coupled receptor family member.
The study used millions of AI-predicted protein structures to identify the protein.
The protein belongs to the 'dark proteome'—proteins whose functions cannot be inferred from sequence similarity.

The Discovery

Lee et al. used predicted protein structures to identify a previously uncharacterized protein. The protein shows structural similarity to G-protein-coupled receptors (GPCRs), a large family of cell surface receptors. Despite this resemblance, the protein exhibits unconventional behavior distinct from typical GPCRs. The authors mined millions of AI-predicted structures to find this evolutionarily conserved protein.

Dark Proteome Context

Many human proteins remain part of the 'dark proteome' because their amino-acid sequences lack similarity to characterized families. This protein's function could not be inferred from sequence alone, placing it in the dark proteome. The study demonstrates how AI-predicted structures can illuminate proteins that sequence-based methods miss.