PlastoGram

PlastoGram predicts the genomic origin and subplastid localization of plastid proteins, including localization to the envelope, stroma, thylakoid membrane, and thylakoid lumen, and predicts thylakoid lumen import pathway when applicable to inform studies of plastid biogenesis and function.

Key Features:

• Manually Curated Dataset: PlastoGram leverages a manually curated dataset of plastid proteins that underpins its localization predictions.
• Ensemble Model Approach: The tool employs an ensemble modeling strategy that integrates multiple predictive algorithms to improve accuracy and address limited dataset sizes and homology reduction.
• Comprehensive Localization Prediction: PlastoGram classifies proteins by encoding origin (nuclear or plastid) and predicts precise localization within envelope, stroma, thylakoid membrane, and thylakoid lumen, including prediction of the thylakoid lumen import pathway.
• Differentiation of Membrane Proteins: The method distinguishes nuclear-encoded inner versus outer plastid membrane proteins to increase localization granularity.

Scientific Applications:

• Proteome annotation: Accurate subplastid localization and origin predictions support annotation of plastid proteomes and assignment of protein function.
• Protein targeting and import studies: Predictions of thylakoid lumen localization and import pathway inform studies of protein targeting and import mechanisms.
• Evolution, biogenesis, and function studies: Combined origin and subplastid localization data support research on plastid evolution, biogenesis, and functional organization.

Methodology:

PlastoGram builds an ensemble model that integrates various predictive algorithms and is trained/evaluated using a manually curated plastid-protein dataset to address limited data sizes and homology reduction.