{
  "id": 13858632,
  "title": "CHORUS: A proteome-scale atlas of membrane interaction interfaces constructed using cross-protein language model embeddings from sequences alone",
  "url": "https://urgent.news/2026/10/11/chorus-a-proteome-scale-atlas-of-membrane-interaction-interfaces",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-11T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.10.04.756538v1?rss=1"
  },
  "original_language": "en",
  "account": "Recent research introduces CHORUS, a deep-learning framework designed to decode membrane protein interaction sites using only sequence data. This novel approach circumvents the need for 3D structures, multiple sequence alignments, and predefined partner identities traditionally required for such analyses.\n\nCHORUS surpasses 18 existing sequence- and structure-based methods in a study involving 4,655 homology-isolated test contexts, achieving an average precision of 0.643. The model's performance improves further to 0.746 when focusing on the lipid-embedded core region of proteins.\n\nOne of CHORUS' standout capabilities is its ability to delineate quaternary contact boundaries across megadalton respiratory supercomplexes, even without sequence alignments, with an impressive precision of 0.984. The researchers have compiled the CHORUS Atlas, an online resource providing residue-resolved interaction landscapes for 80,619 reviewed transmembrane proteins across 4,918 species. This vast dataset spans 33.7 million residues, and is now accessible at https://chorus.atlas.sunnylab.org.\n\nBy offering this extensive, precomputed resource, CHORUS aims to facilitate interactome-scale discovery and allosteric therapeutic design, ultimately advancing our understanding of membrane proteins and their roles in vital cellular processes.",
  "summary": "Membrane proteins drive vital cellular signaling and bioenergetics yet resolving their physical interaction interfaces remains hindered by the paucity of experimental complexes and the breakdown of sequence alignments across hydrophobic bilayers. Here, we present CHORUS, a sequence-only deep-learning framework that decodes interaction sites across membrane proteins without requiring 3D…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}