SWISS-MODEL

SWISS-MODEL generates three-dimensional protein structure models by automated homology modeling to predict protein folds, oligomeric assemblies, and ligand-bound states when experimental structures are unavailable.

Key Features:

• Automated Homology Modeling Pipeline: Automates generation of protein structure models from amino acid sequences using homology modeling.
• SWISS-MODEL Repository (SMR): Contains over 400,000 annotated 3D protein models representing approximately 20% of Swiss-Prot/UniProtKB entries and is updated with new target sequences and template structures.
• Model Quality Estimation: Implements model quality estimation methods such as QMEANDisCo to assess local and global model accuracy.
• Template Library and Selection: Provides a template library annotated with quaternary structure information and essential ligands/co-factors, with templates clustered by sequence similarity for selection.
• Guided Template Selection: Guided template selection enables construction of models in alternative functional states.
• Modeling of Complexes: Extends modeling to homo- and heteromeric complexes, inferring stoichiometry and overall structure from interacting protein sequences.
• ProMod3 Engine: Uses the ProMod3 modeling engine to improve accuracy and efficiency of model building.
• Functional Annotations: Associates functional annotations with generated models to aid biological interpretation.
• Continuous Evaluation: Model accuracy and performance are continuously evaluated using the CAMEO system.

Scientific Applications:

• Structure Prediction: Predicts three-dimensional protein structures to bridge sequence data and experimentally determined structures.
• Drug Discovery: Supports structure-based drug discovery and analysis of protein-ligand interactions.
• Functional Annotation: Aids functional annotation of proteins by providing structural context for sequence-based hypotheses.
• Protein-Protein Interaction Analysis: Models protein-protein interfaces and assemblies to study interaction interfaces and stoichiometry.
• Experimental Design and Interpretation: Supports hypothesis generation, experimental design, and interpretation of biochemical and biophysical data using structural models.

Methodology:

Computational methods explicitly include automated homology modeling, template selection with sequence-similarity clustering, model building with the ProMod3 engine, model quality estimation using QMEANDisCo, and continuous benchmarking with CAMEO.