TatP

TatP predicts twin-arginine (Tat) signal peptides and their potential cleavage sites in bacterial proteins to identify proteins targeted to the periplasmic compartment or extracellular environment via the Tat translocation pathway.

Key Features:

• High accuracy: Correctly classifies 91% of known Tat signal peptides and predicts 84% of annotated cleavage sites.
• Reduction of false positives: Produces fewer false positive predictions compared to simple pattern matching and complementary rule-based methods.
• Discrimination of signal types: Distinguishes Tat signal peptides from cytoplasmic proteins with similar motifs and from Sec signal peptides.
• Hydrophobicity-based artificial neural network: Employs an artificial neural network to separate Tat and Sec signal peptides based on hydrophobicity profiles.
• Data training on known substrates: Trained using sequence data from known Tat substrates to enhance predictive performance.
• Regular expression filtering: Supports filtering of input sequences using Perl syntax regular expressions.
• Cleavage site reporting: Reports potential cleavage sites to inform analyses of protein processing and maturation.

Scientific Applications:

• Tat pathway studies: Identification of Tat substrates for research on the twin-arginine translocation system.
• Protein targeting and translocation analysis: Analysis of signals that direct proteins to the periplasmic compartment or extracellular environment.
• Functional annotation: Distinguishing secreted/periplasmic proteins from cytoplasmic proteins for genome and proteome annotation.
• Microbial physiology and pathogenesis: Investigation of protein localization relevant to microbial physiology and virulence mechanisms.
• Synthetic biology and biotechnology: Selection and validation of signal peptides for engineering secretion or periplasmic localization in bacterial systems.

Methodology:

Sequence data from known Tat substrates were used to train the model; an artificial neural network discriminates Tat versus Sec signal peptides based on hydrophobicity profiles; input sequences can be filtered using Perl syntax regular expressions.