ANDIS

ANDIS implements an Atomic ANgle- and DIStance-dependent statistical potential to assess protein structure quality by balancing native recognition and decoy discrimination.

Key Features:

• Tunable Distance Cutoff Parameter: ANDIS employs a flexible distance cutoff parameter; for distance cutoffs ≤9.0 Å it applies "effective atomic interaction" strategies to improve native recognition, while for cutoffs ≥10 Å it uses a distance-dependent atom-pair potential with a random-walk reference state to enhance decoy discrimination.
• Dual-Aspect Optimization: ANDIS addresses both native recognition and decoy discrimination by adapting its potential based on the chosen distance cutoff.
• Benchmark Performance: Evaluation across 632 structural decoy sets demonstrates that ANDIS outperforms other state-of-the-art potentials in both native recognition and decoy discrimination.

Scientific Applications:

• Protein Structure Modeling: Assessing and ranking candidate protein models to support generation of accurate structural models.
• Model Quality Assessment: Distinguishing native conformations from incorrect decoys to evaluate structural integrity of protein models.

Methodology:

Integration of angle- and distance-dependent interactions into a statistical potential with a tunable distance cutoff, employing effective atomic interaction strategies for ≤9.0 Å and a distance-dependent atom-pair potential with a random-walk reference state for ≥10 Å.