WS-SNPs-and-GO

WS-SNPs-and-GO predicts disease-associated single nucleotide polymorphisms that cause amino acid changes in proteins by integrating sequence, three-dimensional structural information, and functional annotations for pathogenicity assessment.

Key Features:

• Integration of sequence and structural data: Employs a structure-based machine learning approach (SVM-3D) that uses protein sequence, three-dimensional structure, and functional annotation features.
• Training dataset: Trained on 3,342 disease-related mutations and 1,644 neutral polymorphisms derived from 784 protein chains.
• Predictive performance: SVM-3D reports 85% overall accuracy, a correlation coefficient of 0.70, and an AUC of 0.92.
• Improvement over sequence-only methods: Shows a ~3% increase in accuracy and a +0.06 increase in correlation coefficient compared to sequence-based predictors.
• Functional annotation integration: Incorporates Gene Ontology terms and reports 82% scoring efficiency with a Matthews correlation coefficient of 0.63 on annotated nonsynonymous mutations.
• Comparative robustness: Validated across multiple datasets and reported to outperform prior methods including PolyPhen2, with SNPs&GO combining sequence, evolutionary, and functional data.

Scientific Applications:

• Disease susceptibility research: Identification of disease-associated SNPs to characterize genetic markers linked to pathologies.
• Protein function analysis: Assessment of the impact of amino acid substitutions on protein function and molecular mechanisms.
• Drug development and personalized medicine: Prioritization of mutations to inform drug target selection and personalized treatment strategies.

Methodology:

Trains a Support Vector Machine (SVM) on disease-related and neutral polymorphisms using features derived from protein sequence, three-dimensional structure (from the Protein Data Bank), and functional annotations.