PhD-SNP

PhD-SNP predicts the deleteriousness of non-synonymous single nucleotide polymorphisms (nsSNPs) in human proteins to identify variants associated with genetic disease.

Key Features:

• Support Vector Machine Algorithm: Employs support vector machines (SVMs) to process protein sequence information and discriminate disease-related from neutral single-point mutations.
• High Accuracy: Demonstrates prediction accuracy exceeding 74% for determining whether a single point mutation is associated with genetic diseases.
• Comprehensive Dataset: Developed and validated on a dataset comprising 21,185 single-point mutations across 3,587 proteins, with 61% labeled as disease-related.
• Focus on Non-Synonymous SNPs (nsSNPs): Concentrates on nsSNPs that cause amino acid changes in protein sequences.
• Application to Significant Diseases: Addresses mutations relevant to diseases including Alzheimer's, Parkinson's, and Creutzfeldt-Jakob's diseases.

Scientific Applications:

• Disease Prediction and Research: Identifies nsSNPs that may contribute to genetic disorders, aiding research into pathogenic variants.
• Genetic Variation Analysis: Provides insights into how specific amino acid substitutions affect protein function and disease susceptibility.
• Biomedical Informatics: Integrates machine learning with biological sequence data to support genotype–phenotype investigations.

Methodology:

The method trains a support vector machine on a dataset of known nsSNPs (21,185 single-point mutations across 3,587 proteins) and analyzes sequence context and structural implications of each mutation to assess the likelihood that a SNP is deleterious.