{
  "id": 7329423,
  "title": "Assessing multi-eGO Predictions of PDZ2-Peptide Binding across Mutations",
  "url": "https://urgent.news/2026/09/14/assessing-multi-ego-predictions-of-pdz2-peptide-binding-across",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-14T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.10.750618v1?rss=1"
  },
  "original_language": "en",
  "account": "Predicting how mutations impact protein-peptide binding poses significant challenges for molecular simulations due to the computational complexity involved in both conformational sampling and binding kinetics. In this study, researchers sought to determine if multi-eGO, a hybrid transferable/structure-based atomistic-resolution model validated for protein-small molecule interactions, could be adapted for protein-peptide binding without requiring peptide-specific retraining. Employing the PDZ2 domain of protein tyrosine phosphatase basophil-like in complex with the peptide EQVTAV as a benchmark, the team first demonstrated that multi-eGO effectively captures the structural dynamics of PDZ2 and the equilibrium binding thermodynamics of the wild-type complex. Remarkably, the model accelerated both binding and unbinding processes compared to experimental measurements while retaining the correct equilibrium dissociation constant. Subsequently, the researchers introduced conservative mutations in both PDZ2 and the peptide and evaluated their effects on binding using multi-eGO, avoiding the need for further computationally expensive training. The results showed strong agreement between multi-eGO and experimental observations regarding changes in equilibrium dissociation constants across PDZ2 mutants, albeit with moderate agreement when the peptide was also mutated. However, the individual association and dissociation rate constants exhibited weaker agreement with experimental data. Collectively, these findings indicate that multi-eGO's simplified energy landscape hinders the precise prediction of absolute kinetics but effectively preserves thermodynamic information crucial for relative binding affinity. Consequently, multi-eGO emerges as a computationally efficient approach for analyzing protein-peptide recognition and predicting and ranking the impact of conservative mutations on binding affinity.",
  "summary": "Accurately predicting how mutations alter protein-peptide binding remains challenging for molecular simulations because both conformational sampling and binding kinetics are computationally demanding. Here, we investigate whether multi-eGO, a hybrid transferable/structure-based atomistic-resolution model previously developed and validated for protein-small molecule interactions, can be…",
  "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."
}