ProQ3D

ProQ3D assesses protein model quality by estimating local (per-residue) errors to inform model refinement and improve outcomes in molecular replacement.

Key Features:

• Local Error Estimation: Provides detailed local (per-residue) error estimates for protein structural models to identify regions requiring correction.
• Enhanced Molecular Replacement (MR): Uses local error estimates to adjust B factors and has been reported to increase log-likelihood gain (LLG) scores on average by over 50%.
• Comparative Performance: In an evaluation of 431 homology models targeting difficult MR cases, 209/431 (48.5%) models attained LLG>50 with ProQ3D error estimates versus 40.6% with ProQ2 and 17.2% without error estimation.

Scientific Applications:

• Molecular replacement: Improving MR success by adjusting B factors based on ProQ3D local error estimates to increase LLG and yield more successful MR solutions for challenging targets.
• Model refinement and error localization: Identifying specific regions of protein models that require refinement to enhance structural prediction accuracy.

Methodology:

Analyzes input protein models to estimate local (per-residue) errors, and uses those error estimates to adjust model parameters such as B factors, leading to improved log-likelihood gain (LLG) in molecular replacement.