PhD-SNPg

PhD-SNPg predicts the pathogenicity of single nucleotide variants (SNVs) in both coding and non-coding regions of the human genome by using sequence-based features in a binary classification framework.

Key Features:

• Binary classifier: Implements a binary classification to distinguish pathogenic versus benign SNVs.
• Machine learning: Uses a machine-learning approach based solely on sequence-derived features.
• Sequence-based features: Relies exclusively on sequence information rather than extensive pre-calculated annotations.
• Sequence conservation and functional features: Integrates sequence conservation and functional features into the predictive model.
• Coding and non-coding evaluation: Predicts variant effects in both coding and non-coding regions of the human genome.
• Comparative performance: Reports performance comparable to or exceeding established methods such as CADD and FATHMM.

Scientific Applications:

• Pathogenicity prediction: Assessment of the pathogenic potential of single nucleotide variants in human genetic studies.
• Variant annotation: Annotation of SNVs using conservation and functional sequence features for interpretation of genetic variation.
• Method benchmarking: Serves as a benchmark for evaluating and developing other variant effect prediction methods.

Methodology:

PhD-SNPg employs a machine-learning binary classifier that uses only sequence-based features, including sequence conservation and functional features, to predict SNV pathogenicity.