fidosnp

fidosnp predicts the functional impact of single nucleotide variants (SNVs) in the canine genome to support interpretation of genomic variation for veterinary and comparative genomics.

Key Features:

• Species-Specific Design: Tailored for the dog (canine) genome to address non-human genomic variant interpretation.
• Binary Classification: Classifies SNVs into pathogenic versus benign categories.
• Algorithm: Implements a Gradient Boosting machine learning model for classification.
• Variant Scope: Evaluates both coding and non-coding genomic regions.
• Input Features: Uses sequence-derived features to inform predictions.
• Performance Metrics: Validated on annotated variants from the OMIA database with reported accuracy 88%, MCC 0.77, and AUC-ROC 0.91.
• Runtime: Provides rapid assessments, reporting predictions within seconds.

Scientific Applications:

• Veterinary Genetics: Prioritizes candidate SNVs for association with canine diseases and traits.
• Disease Mechanism Investigation: Guides identification of potentially pathogenic variants to study molecular disease mechanisms.
• Canine Personalized Medicine: Supports interpretation of individual dog genomes for clinical or breeding decisions.
• Conservation and Breeding Programs: Assists in genomic analyses relevant to conservation biology and selective breeding strategies.
• Therapeutic Development: Aids in selecting variant targets relevant to development of targeted therapies for canine diseases.

Methodology:

Uses a Gradient Boosting machine learning approach trained on sequence data to learn patterns distinguishing pathogenic and benign SNVs.