Hansa

Hansa predicts the phenotypic impact of nonsynonymous single nucleotide polymorphisms (nsSNPs) by classifying amino-acid–altering variants as disease-associated (pathogenic) or benign (neutral).

Key Features:

• Classification algorithm: Implements a support vector machine (SVM)-based classifier to distinguish pathogenic and neutral nsSNPs.
• Discriminatory features: Leverages 10 Neutral-Disease Mis-Sense Mutation Discriminatory (NDMSMD) features engineered to enhance prediction accuracy.
• Variant scope: Targets nonsynonymous single nucleotide polymorphisms that result in amino acid changes in proteins.
• Benchmarking: Evaluated with fivefold cross-validation on the HumVar dataset, reporting a true positive rate (TPR) of 82% at a false positive rate (FPR) of 20%.
• Independent validation: Performance corroborated on independent datasets comprising well-characterized disease and neutral mutations, showing approximately 10% improvement over the prior best-known method.

Scientific Applications:

• nsSNP effect prediction: Assigns pathogenic versus neutral status to nonsynonymous variants to inform phenotypic impact assessments.
• Variant prioritization: Ranks candidate disease-associated mutations for genetic studies and clinical interpretation.
• Genetic research and personalized medicine: Supports identification of deleterious mutations relevant to human health and personalized therapeutic decision-making.

Methodology:

Hansa trains a support vector machine on 10 NDMSMD features to classify nsSNPs; evaluation used fivefold cross-validation on the HumVar benchmark (TPR 82% at FPR 20%) and validation on independent datasets of characterized disease and neutral mutations.