EVAcon

EVAcon evaluates inter-residue contact predictions from protein structure prediction servers by comparing those predictions to newly released experimental structures in the Protein Data Bank (PDB) to monitor and compare server performance.

Key Features:

• Automated evaluation: Automates assessment of contact predictions from nine monitored servers (four specialized in contact prediction and five general structure prediction servers) by comparing predictions to experimental structures.
• Continuous monitoring of PDB releases: Tracks newly determined experimental structures deposited into the Protein Data Bank (PDB) at an observed rate of approximately 5–50 entries per week for ongoing evaluation.
• Dynamic comparison using common protein subsets: Uses a system of common protein subsets to enable standardized, comparable evaluations across servers.
• CASP-aligned criteria: Applies standardized evaluation criteria that align with Critical Assessment of Structure Prediction (CASP) assessments for consistency with community benchmarks.
• Integration with EVA system: Operates as part of the broader EVA continuous evaluation framework to support automated, ongoing server comparisons.

Scientific Applications:

• Server benchmarking: Identifies and compares the accuracy of inter-residue contact and general protein structure prediction servers using experimentally validated structures.
• Method development assessment: Provides continuous performance feedback for developers of contact-prediction algorithms and structure prediction methods.
• Structural biology and drug discovery: Informs studies of protein folding and function and supports applications relevant to drug discovery and therapeutic development by highlighting reliable prediction resources.

Methodology:

Automated comparison of server-predicted inter-residue contacts to experimental PDB structures using common protein subsets and evaluation criteria aligned with CASP, monitoring nine servers.