LOCtarget

LOCtarget predicts and annotates the sub-cellular localization of eukaryotic and prokaryotic proteins to support structural genomics and protein annotation.

Key Features:

• Integration of Multiple Prediction Methods: Combines four complementary methods to generate localization predictions.
• PredictNLS: Identifies known nuclear localization signals within protein sequences.
• LOChom: Transfers experimental localization annotations by homology.
• LOCkey: Infers localization from automatic text analysis of SWISS-PROT keywords.
• LOCnet: Uses a neural network for de novo localization prediction.
• SignalP Integration: Incorporates SignalP to detect signal peptides that target proteins to specific compartments.
• Confidence-Based Prediction: Selects the final localization based on the method yielding the highest confidence score.
• Database of Predictions: Contains weekly-updated precomputed localization predictions for all eukaryotic proteins from TargetDB.

Scientific Applications:

• Structural Genomics: Supports target selection and interpretation in structural genomics through localization data.
• Protein Annotation: Aids annotation of newly sequenced proteins with predicted sub-cellular localization.
• Protein Trafficking Studies: Facilitates study of protein trafficking pathways by providing localization hypotheses.
• Experimental Design: Informs design of experiments involving protein targeting and modification.

Methodology:

Uses PredictNLS to identify nuclear localization signals; LOChom for homology-based transfer of experimental annotations; LOCkey for automatic text analysis of SWISS-PROT keywords; LOCnet as a neural network for de novo prediction; incorporates SignalP for signal peptide detection; selects the final localization by the highest confidence score; and maintains a weekly-updated database of precomputed predictions for eukaryotic proteins from TargetDB.