PredictSNP

PredictSNP predicts the effects of single nucleotide variants (SNVs) on protein function by integrating outputs from multiple variant-effect predictors into a consensus classifier.

Key Features:

• Consensus Classifier: Combines predictions from MAPP, nsSNPAnalyzer, PANTHER, PhD-SNP, PolyPhen-2, and SIFT into a unified score for SNV impact on protein function.
• Three Independent Datasets: Constructs three independent datasets to eliminate duplicities and inconsistencies for unbiased evaluation.
• Extensive Benchmarking: Evaluates eight established prediction tools using a benchmarking set of over 43,000 mutations.
• Rich Annotations: Integrates mutation annotations from the Protein Mutant Database and UniProt.

Scientific Applications:

• Variant Prioritization: Prioritizes SNVs for experimental characterization based on predicted functional impact.
• Disease Mechanism Studies: Supports studies of the molecular basis of genetic diseases by providing predicted effects of missense mutations on protein function.

Methodology:

Rigorous dataset curation to remove overlaps and inconsistencies and construction of three independent datasets, integration of predictions from MAPP, nsSNPAnalyzer, PANTHER, PhD-SNP, PolyPhen-2, and SIFT into a consensus classifier, and benchmarking on a set of over 43,000 mutations to assess eight established prediction tools.