SigniSite

SigniSite identifies residue-level genotype-phenotype correlations within protein multiple sequence alignments using quantitative (real-number) phenotypic data.

Key Features:

• Subgroup-Free Analysis: Performs correlation analysis without predefined subgroups or binary classification, enabling use of continuous phenotype values.
• Input Requirements: Accepts protein multiple sequence alignments with each sequence associated with a real-number phenotype quantification.
• Statistical Identification: Identifies amino acid residues significantly correlated with the phenotype using statistical methods.
• Visualization Outputs: Produces sequence logos that represent residue–phenotype association strength and heat-maps that highlight 'hot' and 'cold' regions of association.
• Benchmarking and Performance: Benchmarked against SPEER using human immunodeficiency virus protease–inhibitor genotype-phenotype data from the Stanford University HIV Drug Resistance Database and protein families with experimentally annotated specificity-determining positions (SDPs), and shown to outperform SPEER.

Scientific Applications:

• Mutation Functional Analysis: Pinpoints specific amino acid residues associated with phenotypic changes to investigate the functional impact of mutations.
• Disease Mechanisms: Supports analysis of molecular mechanisms of disease, exemplified by drug resistance in HIV.
• Therapeutic Target Identification: Aids identification of residue-level targets for therapeutic intervention based on genotype-phenotype correlations.

Methodology:

Accepts protein multiple sequence alignments paired with real-number phenotypes, performs subgroup-free statistical tests to identify residues significantly correlated with phenotype, and generates sequence logo and heat-map visualizations.