MultiLoc2

MultiLoc2 predicts protein subcellular localization for animals, plants, and fungi to support proteomics, drug target discovery, and systems biology.

Key Features:

• Integration of phylogenetic profiles: Incorporates phylogenetic profiles to capture evolutionary relationships among proteins.
• Use of Gene Ontology (GO) terms: Employs GO terms as structured functional annotations to inform localization predictions.
• Two prediction versions: Provides a Globular Protein Version that predicts up to five subcellular localizations and a Comprehensive Eukaryotic Localization Version that covers all eleven main eukaryotic subcellular localizations.
• Training datasets: Trained on two distinct datasets to develop and validate the prediction models.
• Benchmark performance: Demonstrated superior performance relative to other methods for animal and plant proteins and comparable performance for fungal proteins, with improved accuracy when using an extended dataset covering all eleven main eukaryotic localizations.

Scientific Applications:

• Proteomics: Enables assignment of subcellular localization to proteins to aid proteome annotation and interpretation.
• Drug target discovery: Provides localization context to inform identification and prioritization of drug targets.
• Systems biology: Supplies localization constraints for proteins to support modeling of cellular processes.

Methodology:

Integrates phylogenetic profiles and Gene Ontology (GO) terms; models were trained on two distinct datasets and benchmarked against other localization prediction methods.