FoldScript

FoldScript synthesizes primary to quaternary structural information from sets of AI-generated protein models to enable rational selection of the most accurate 3D protein structure.

Key Features:

• Multi-model analysis: Gathers and synthesizes primary to quaternary structure information from up to 25 AI-generated models.
• Visual comparison: Presents analysis as a comprehensible flat figure and mosaics of 3D structures for efficient visual inspection of structural differences.
• Contact identification: Identifies models with known intermolecular contacts.
• Cross-predictor evaluation: Evaluates and compares AlphaFold 2 and AlphaFold 3 models.

Scientific Applications:

• Oligomeric model selection: Supports selection of optimal oligomeric models where the highest scoring solution may not be the best, particularly when incorporating experimental knowledge.
• AI model benchmarking: Facilitates comparison of results from different AI predictors for improved structure prediction outcomes.
• Structural biology decision support: Assists both structural biologists and non-specialists in rational selection of accurate models from AI-driven protein modelling processes.

Methodology:

FoldScript analyzes structural information from sets of AI-generated models, with specific computational steps not explicitly detailed in the sources beyond synthesis of primary through quaternary structural data and identification of models containing known intermolecular contacts.