Jpred 3

Jpred 3 predicts protein secondary structure and residue solvent accessibility to assign three-state labels (alpha helix, beta strand, coil) for protein sequences and multiple sequence alignments.

Key Features:

• Jnet algorithm and accuracy: Implements the Jnet algorithm with a reported 81.5% three-state secondary structure prediction accuracy.
• Prediction targets: Produces per-residue three-state secondary structure assignments (alpha helix, beta strand, coil) and solvent accessibility predictions.
• Input types and profile generation: Accepts single protein sequences or multiple sequence alignments and generates alignment profiles to inform predictions.
• Database integration: Synchronizes reference data with SCOP and UniProt via a software pipeline to maintain up-to-date mappings.
• Output formats: Provides multiple output formats including colored HTML, plain text, PostScript, PDF, and Jalview-compatible files.

Scientific Applications:

• Structural biology: Supplies residue-level secondary structure information for interpretation of protein folding, motifs, and domain organization.
• Protein function and interaction analysis: Informs analyses of protein function, stability, and potential interaction interfaces through secondary structure and solvent accessibility data.
• Annotation and curation: Facilitates annotation of protein sequences and integration with SCOP and UniProt reference data for database curation.

Methodology:

Uses the Jnet algorithm to predict three-state secondary structure and solvent accessibility from single sequences or multiple sequence alignments by generating alignment profiles and synchronizing reference data with SCOP and UniProt via a software pipeline.