Parepro

Parepro predicts the impact of non-synonymous single nucleotide polymorphisms (nsSNPs) on protein function to identify amino acid replacements that are deleterious or neutral for genetic-disease research.

Key Features:

• Support Vector Machine (SVM) approach: Parepro employs a support vector machine framework to analyze nsSNPs and predict their effects on protein function.
• Training datasets: The model was trained using the HumVar and NewHumVar datasets obtained from the PhD-SNP server.
• Performance metrics: On the HumVar dataset Parepro reported a Matthews correlation coefficient (MCC) of 50% and an overall accuracy (Q2) of 76%, with similar high-performance results on NewHumVar.
• Comparative evaluation: Performance was compared against PolyPhen, SIFT, and HybridMeth and reported as superior based on the stated metrics.
• Cross-validation testing: A 20-fold cross-validation procedure was employed on the HumVar dataset to assess prediction reliability.

Scientific Applications:

• Medical genetics: Identification and prioritization of nsSNPs that may contribute to inherited human disease phenotypes by predicting deleterious versus neutral amino acid substitutions.
• Genotype-to-phenotype studies: Analysis of the molecular consequences of amino acid changes to inform studies on the molecular basis of genetic disorders and potential therapeutic targets.

Methodology:

Parepro trains an SVM model using the HumVar and NewHumVar datasets from the PhD-SNP server; a 20-fold cross-validation was performed on HumVar; comparative performance analysis was conducted against PolyPhen, SIFT, and HybridMeth.