{
  "id": 3125916,
  "title": "MG2Act: A Mechanism-Inspired Sequential Attention Framework for Molecular Glue Degradation Prediction",
  "url": "https://urgent.news/2026/08/17/mg2act-a-mechanism-inspired-sequential-attention-framework-for",
  "topic": "science",
  "section": "Science",
  "published": "2026-08-17T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.08.08.739980v1?rss=1"
  },
  "original_language": "en",
  "account": "Molecular glue degraders function by promoting the proximity between an E3 ligase and a substrate protein. In most characterized degradative glues, a small molecule initially interacts with the E3, which modifies its substrate-recognition surface and allows recruitment of a compatible neo-substrate. This sequential process is seldom explicitly modeled in computational frameworks, which usually merge molecule, E3, and target representations concurrently. To address this limitation, we introduce MG2Act, a structure-agnostic framework that embodies this two-step mechanism using sequential cross-attention, drawing inspiration from CRBN-mediated degradation as a model system. Our research begins with a comprehensive benchmark of 1,207 pairs spanning 47 targets, from which a refined dataset of 1,159 pairs was curated by eliminating rare targets. Upon training on this refined dataset, MG2Act demonstrated superior performance compared to existing machine learning models. Notably, MG2Act exhibited robust generalization abilities when subjected to strict redundancy-filtering conditions and demonstrated coherent responses to mechanism-based perturbations. Furthermore, our framework was utilized to identify nanomolar degraders for the proteins IKZF1, CK1, and CDK4, encompassing the non-classical IMiD-core CDK4 degrader SWC-202.",
  "summary": "Molecular glue degraders act by inducing productive proximity between an E3 ligase and a substrate protein. For most characterized degradative glues, a small molecule first engages the E3, conditions its substrate-recognition surface, and only then enables recruitment of a compatible neo-substrate. This directionality is rarely encoded explicitly in computational models, which typically fuse…",
  "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."
}