NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·2h ago
AI Research
arXiv CS.AI·May 27·5 min read
Co-folding model guided by structural proteomics
TL;DR
Predicting how proteins fold into 3D structures remains challenging when structural data is limited or incomplete. Researchers developed a co-folding model that leverages structural proteomics—large-scale experimental protein structure data—to guide and improve folding predictions.
✦ Why It Matters
Engineers can integrate experimental proteomics data into folding pipelines to achieve higher-confidence structure predictions for drug discovery and protein engineering.
Key Takeaways
- Utilize AIMS-Fold for improved protein complex structure predictions.
- Incorporate XL-MS and HDX-MS data for better accuracy.
- Leverage pretrained diffusion models in your research.
- Explore structure-based drug design using this integrative approach.
How It Works
AIMS-Fold combines structural proteomics data with pretrained diffusion models to guide the generative sampling of protein structures. It uses differentiable physical potentials derived from XL-MS spatial restraints and HDX-MS solvent accessibility profiles to actively steer the model's predictions, enhancing the accuracy of conformational state predictions for protein complexes.