14-3-3-Pred

14-3-3-Pred predicts 14-3-3 protein binding sites on phosphoproteins by identifying motifs surrounding phosphorylated serine (Ser) and threonine (Thr) residues using machine learning classifiers.

Key Features:

• Multi-Classifier Prediction Framework: Integrates Artificial Neural Networks (ANN), Position-Specific Scoring Matrix (PSSM), and Support Vector Machines (SVM) to predict 14-3-3-binding motifs.
• Motif Window Analysis: Analyzes sequence motifs within a −6 to +4 amino acid window surrounding phosphorylation sites to detect potential 14-3-3-binding regions.
• Predictive Performance Metrics: Achieves Matthews correlation coefficient (MCC) values up to 0.60 and demonstrates improved prediction accuracy compared with Scansite and ELM in blind testing.
• Proteome-Scale Scanning: Supports analysis of the entire human proteome to identify potential 14-3-3-binding phosphosites.
• Experimentally Validated Predictions: Includes validated predictions in proteins such as FAM122A and FAM122B.

Scientific Applications:

• Identification of 14-3-3 Interaction Sites: Enables discovery of novel phosphoprotein binding motifs interacting with 14-3-3 proteins.
• Prioritization of Experimental Targets: Supports selection of candidate proteins from high-throughput datasets for experimental validation.
• Interactome Analysis: Contributes predicted binding sites to the ANIA (ANnotation and Integrated Analysis of the 14-3-3 interactome) database for analysis of the 14-3-3 interaction network.

Methodology:

The tool trains Artificial Neural Network (ANN), Position-Specific Scoring Matrix (PSSM), and Support Vector Machine (SVM) classifiers on curated datasets of experimentally validated 14-3-3-binding motifs and evaluates sequence motifs within a −6 to +4 window around phosphorylated Ser and Thr residues to identify binding sites.