Struct2Net

Struct2Net predicts interactions between arbitrary protein pairs using a structure-based approach that infers protein–protein interactions from protein sequence information and structural data.

Key Features:

• Structure-based prediction: Leverages protein structural data to predict PPIs beyond traditional homology modeling techniques.
• Independence from functional genomic data: Operates without relying on Gene Ontology (GO) annotations, gene expression profiles, or cellular localization information.
• Sequence-only input: Requires only protein sequence information as input for prediction.
• Pre-computed model-organism predictions: Provides pre-computed predictions for Drosophila melanogaster (fly), Homo sapiens (human), and Saccharomyces cerevisiae (yeast).
• Orthology-based approximation: Offers a quick, approximate option for non-model species by mapping orthologs to pre-computed data.
• Comprehensive computation: Offers a full computation mode for non-model species to generate more detailed predictions.
• Query modes: Supports multiple querying modes including retrieval of pre-computed results, orthology-based mapping, and comprehensive de novo computation.

Scientific Applications:

• Experimental prioritization: Prioritizes candidate PPIs to guide experimental validation efforts.
• Drug discovery and design: Informs drug design and target selection by providing structure-based interaction hypotheses.
• Interactome expansion: Fills gaps in experimental coverage of the PPI interactome where experimental data are limited.

Methodology:

Structure-based inference using protein structural data with sequence-only input; predictions pre-computed for D. melanogaster, H. sapiens, and S. cerevisiae; for other species supports orthology-based mapping to pre-computed data or comprehensive de novo computation.