BTXpred

BTXpred predicts bacterial toxins and classifies their types and functions from primary amino acid sequences to support microbiological and toxicological analyses.

Key Features:

• Prediction of Bacterial Toxins: Support vector machine (SVM) models using amino acid and dipeptide composition predict bacterial toxins from primary amino acid sequences with reported accuracies of 96.07% (amino acid composition) and 92.50% (dipeptide composition).
• Discrimination Between Toxin Types: SVM-based modules differentiate endotoxins and exotoxins using amino acid and dipeptide composition with reported accuracies of 95.71% and 92.86%, respectively.
• Classification of Exotoxins: Classification of exotoxin subtypes (adenylate cyclase activators, guanylate cyclase activators, neurotoxins) using hidden Markov models (HMM), PSI-BLAST, and their combination with reported accuracies of 95.75% (HMM), 97.87% (PSI-BLAST), and 100% (combination).

Scientific Applications:

• Toxin Discovery: Assisting identification of novel bacterial toxins from genomic or proteomic sequence data.
• Pathogenicity Studies: Enabling analysis of bacterial pathogenesis through toxin type and function classification.
• Drug Development: Informing design of inhibitors targeting specific bacterial toxin classes.

Methodology:

SVM models trained on amino acid and dipeptide composition; SVM-based modules for endotoxin-versus-exotoxin discrimination; exotoxin classification using HMM, PSI-BLAST, and their combination; trained and tested on a non-redundant dataset of 150 bacterial toxins (77 exotoxins, 73 endotoxins).