CryoRes

CryoRes estimates local resolution in cryo-electron microscopy (cryo-EM) density maps using a deep-learning framework to provide per-voxel resolution assessments for structural validation and interpretation.

Key Features:

• Deep Learning Approach: Employs a deep-learning framework trained on 1,174 experimental maps to learn resolution-aware patterns in density map voxels.
• Single-Map Input: Operates on a single final cryo-EM density map without requiring half maps or manual parameter adjustments.
• Performance: Achieves an average root mean square error (RMSE) of 2.26 Å compared with the FSC-based method blocres.
• Molecular Mask Generation: Automatically generates molecular masks with a reported 12.12% accuracy improvement over ResMap.
• Speed and Automation: Performs analysis rapidly and without parameter tuning or manual intervention.
• Versatility: Applicable to both standard cryo-EM density maps and cryo-EM subtomogram datasets.

Scientific Applications:

• Quality Assessment of Cryo-EM Maps: Provides local resolution estimates to identify heterogeneous regions within density maps that affect map interpretation.
• Structure Determination Support: Guides refinement and validation of macromolecular structures by indicating regions of differing local resolution.
• Subtomogram Analysis: Extends local resolution estimation workflows to cryo-EM subtomogram datasets.

Methodology:

Uses a deep-learning model trained on 1,174 experimental cryo-EM maps to predict per-voxel local resolution and to generate molecular masks from single-map inputs; performance was evaluated by RMSE against the FSC-based blocres method.