struct-predictor

What problem does it solve?

Many researchers and clinicians need accurate, reproducible 3D protein models and comparisons to experimental structures without sending sensitive data to the cloud; this Skill enables local prediction, retrieval, and quantitative comparison of protein structures to support variant interpretation, model validation, and visualization.

Core Features & Use Cases

• Structure Prediction: Run local predictors such as AlphaFold (ColabFold), Boltz, or Chai to generate 3D models from amino-acid sequences.
• PDB Retrieval & Comparison: Fetch experimental structures from the PDB and compute RMSD and TM-score against predictions.
• Confidence Mapping & Visualization: Produce pLDDT and PAE confidence maps and generate 3D renders for reports or interactive inspection.
• Report Generation: Compile reproducible markdown reports containing visualizations, confidence plots, and comparison tables for publication or review.

Quick Start

Predict the 3D structure of the protein sequence MVLSPADKTNVKAA using ColabFold locally and compare the result to PDB ID 1A3N.