PremPS

PremPS predicts the effects of single amino-acid (missense) mutations on protein stability and pathogenicity using evolutionary and structure-based features to inform molecular interpretation of variants.

Key Features:

• Predictive Accuracy: Demonstrates improved accuracy over previous methods for estimating effects of mutations that increase protein stability by using ten evolutionary- and structure-based features parameterized on 5,000 mutations.
• Pathogenicity Prediction: Predicts pathogenicity of missense mutations using an experimental dataset composed of 2,000 non-neutral and neutral mutations and exhibits superior performance compared to other computational methods.
• Feature Utilization: Employs ten distinct features grouped into six categories, with evolutionary conservation of mutation sites identified as particularly crucial.
• Comparative Performance: Outperforms 25 other computational methods across multiple test sets, indicating consistent robustness.

Scientific Applications:

• Functional Variant Identification: Identifies functionally important protein variants by predicting mutation impacts on stability and function.
• Molecular Mechanism Elucidation: Reveals how specific missense mutations influence protein stability and function to help elucidate underlying molecular mechanisms.
• Protein Design: Informs design of proteins with desired stability characteristics by predicting stabilizing and destabilizing mutation effects.

Methodology:

Requires a protein 3D structure as input and uses ten evolutionary- and structure-based features (with emphasis on evolutionary conservation) parameterized on 5,000 mutations; pathogenicity assessment used an experimental dataset of 2,000 non-neutral and neutral mutations and performance was compared against 25 other methods.