WoLF PSORT

WoLF PSORT predicts protein subcellular localization from amino acid sequences to support functional annotation and the study of protein sorting mechanisms.

Key Features:

• Numerical Localization Feature Conversion: Transforms protein sequences into numerical features by analyzing sorting signals, amino acid composition, and functional motifs such as DNA-binding regions.
• K-Nearest Neighbor Classifier: Classifies proteins using a k-nearest neighbor (k-NN) algorithm that compares query sequence features to proteins with known subcellular localizations.
• Evidence Presentation: Compiles evidence supporting each prediction, including lists of proteins with similar localization features and tables of individual localization features.
• Sequence Alignments and Database References: Produces sequence alignments between query proteins and proteins with known localizations and references UniProt and Gene Ontology annotations.

Scientific Applications:

• Functional Annotation: Infers potential functions of uncharacterized proteins by predicting their subcellular locations.
• Protein Sorting Mechanisms: Supports investigation of mechanisms that direct proteins to specific cellular compartments.
• Comparative Genomics and Evolutionary Studies: Enables comparison of protein localization across species to study evolutionary conservation and divergence.

Methodology:

Protein sequences are analyzed for sorting signals, amino acid composition, and functional motifs and converted into numerical features, which are classified by a k-nearest neighbor algorithm against a database of proteins with known localizations; sequence alignments are generated and evidence (lists and tables) is compiled with references to UniProt and Gene Ontology annotations.