{
  "id": 3776788,
  "title": "WaterFlow: Prediction of Ordered Water Molecule Positions on Protein Structures",
  "url": "https://urgent.news/2026/08/27/waterflow-prediction-of-ordered-water-molecule-positions-on-protein",
  "topic": "science",
  "section": "Science",
  "published": "2026-08-27T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.08.26.747373v1?rss=1"
  },
  "original_language": "en",
  "account": "Ordered water molecules play a crucial role in various protein functions, such as stability, ligand binding, and catalysis. Accurately predicting their positions with sub-angstrom precision could greatly benefit protein design, binding affinity prediction, and automated model building in techniques like X-ray crystallography and cryo-EM. However, current water molecule prediction methods are not as advanced as protein structure predictions.\n\nTo address this gap, the researchers have developed WaterFlow, a novel generator model and confidence model for predicting the positions of ordered water molecules in protein structures. Extensive testing has shown that WaterFlow outperforms existing state-of-the-art models at every precision level. WaterFlow accurately predicts ground truth modeled water molecules, even those located near protein-ligand interactions and on predicted structures. Moreover, WaterFlow predictions align well with experimental data, indicating its potential utility in both prediction and modeling of water molecules.\n\nBy incorporating this improved model, the researchers aim to tackle the data constraint in water molecule prediction. They map the Pareto front of achievable accuracy in water molecule prediction, alongside analysis of different training data schemas. This analysis reveals that the diversity of high-quality structures is currently limiting the possible results. Ultimately, WaterFlow predicts ordered water molecules to serve as a solvent module for structure-based drug design and to aid in water molecule placement during crystallographic refinement.",
  "summary": "Ordered water molecules mediate many protein functions, including stability, ligand binding, and catalysis. Predicting their positions with sub-angstrom accuracy would support protein design, binding affinity prediction, and automated model building in X-ray crystallography and cryo-EM. However, water molecule prediction lags behind protein and other molecule structure predictions. Here, we…",
  "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."
}