{
  "id": 17736026,
  "title": "Differentiable Spin-Center Diffusion Model Enables DEER-Guided Generative Pipeline of Weighted Protein Ensembles",
  "url": "https://urgent.news/2026/11/01/differentiable-spin-center-diffusion-model-enables-deer-guided",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-01T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.10.29.762916v1?rss=1"
  },
  "original_language": "en",
  "account": "Radiolabeled proteins provide a window into their dynamic behavior through distance distributions, known as P(r) profiles, generated using double electron-electron resonance (DEER) spectroscopy. However, the inherent uncertainty in these profiles, stemming from both conformational heterogeneity and rotameric flexibility of spin labels, has hindered a direct translation of P(r) into representative protein ensembles. To address this challenge, researchers have developed a novel differentiable model called spin-center diffusion, which introduces a tunable rotameric width to the spin label representation.\n\nBy integrating this SC diffusion model with a frozen Boltz-2 structure generator, researchers have created a powerful pipeline that leverages the full DEER experimental P(r) to infer the ensemble of protein conformations at inference time. This approach has been rigorously tested on synthetic P(r) profiles representing known compositions, successfully recovering the underlying ensemble populations, including those with bimodal distributions.\n\nThe pipeline's versatility is demonstrated through its ability to model ensembles of intrinsically disordered proteins, such as transporters PfMATE, LmrP, MhsT, and P-glycoprotein, under varying biochemical conditions. By incorporating the differentiable spin label model into structure generation, the approach provides a novel interpretation of DEER distributions as weighted conformational ensembles, offering a thermodynamically grounded framework for generative protein structure prediction methods.",
  "summary": "Biological function entails the interconversion of a protein between distinct conformations, an equilibrium modulated by ligands and other physicochemical conditions. Double electron-electron resonance (DEER) spectroscopy captures these ensembles as distance distributions, P(r), between spin label pairs. However, the width of P(r) convolves protein conformational heterogeneity with the rotameric…",
  "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."
}