STINGAllo

STINGAllo predicts allosteric site-forming residues in proteins using a residue-centric machine-learning model and 54 internal protein nanoenvironment descriptors at single-residue resolution.

Key Features:

• Residue-centric prediction: Predicts allosteric site-forming residues independently of surface geometry, including concave pockets, flat surfaces, and cryptic regions.
• Internal protein nanoenvironment descriptors: Integrates hydrophobic interaction networks, local density, graph connectivity, and a "sponge effect" metric.
• Input flexibility: Accepts protein structures via PDB identifiers or custom uploads.
• Interactive 3D visualization: Provides interactive 3D visualization of predicted pockets.
• RESTful API: Supports integration into computational pipelines through a RESTful application programming interface.

Scientific Applications:

• Allosteric site identification: Detects allosteric sites in proteins including those outside surface invaginations missed by traditional pocket-based predictors.
• Allosteric drug discovery: Accelerates identification of allosteric targets for therapeutic intervention.
• Protein regulation studies: Deepens understanding of allosteric regulation and protein function modulation.
• Large-scale proteome analysis: Enables high-throughput screening of proteins for allosteric site presence across structural databases.

Methodology:

STINGAllo uses a residue-centric machine-learning model trained on 54 optimized internal protein nanoenvironment descriptors including hydrophobic interaction networks, local density, graph connectivity, and a "sponge effect" metric.