Deep-PLA

Deep-PLA predicts enzyme-specific lysine acetylation sites on proteins to characterize HAT and HDAC regulatory mechanisms.

Key Features:

• Enzyme-specific prediction: Predicts lysine acetylation events specific to histone acetyltransferases (HATs) and histone deacetylases (HDACs).
• Deep neural network: Employs a deep learning model for acetylation-site prediction.
• Curated experimental dataset: Trains on extensively curated experimentally identified substrates and modification sites from the literature.
• Sequence feature integration: Integrates diverse protein sequence features as input to the predictive model.
• Particle swarm optimization: Optimizes model performance using particle swarm optimization techniques.
• PPI-network visualization: Visualizes enzyme-specific acetylation regulatory relationships through protein-protein interaction networks.
• Cross-cancer mutation analysis: Examines acetylation-associated mutations across cancers to assess disruptions in acetylation regulation.

Scientific Applications:

• Regulatory mechanism elucidation: Elucidates HAT/HDAC regulatory mechanisms underlying protein lysine acetylation across biological processes.
• Mutation impact assessment: Identifies potential disruptions of acetylation regulation caused by acetylation-associated mutations.
• Cancer signaling and prognosis research: Provides insights into how altered acetylation and associated mutations modulate cancer signaling pathways and inform prognosis and treatment studies.

Methodology:

A deep neural network is trained on curated experimentally identified substrates and modification sites using integrated protein sequence features and optimized with particle swarm optimization; outputs include enzyme-specific acetylation predictions, protein-protein interaction network visualizations, and cross-cancer analyses of acetylation-associated mutations.