Cysteine pattern barcoding-based dataset filtration enhances the machine learning-assisted interpretation of Conus...

Area of Science:

• Bioprospecting and Drug Discovery
• Bioinformatics and Computational Biology
• Pharmacology

Background:

• Cone snail venom is a rich source of bioactive peptides with significant therapeutic potential.
• Understanding cysteine (Cys) patterns and frameworks in these peptides is crucial for drug development.
• Existing methods for analyzing venom peptide potential can be enhanced with advanced computational approaches.

Purpose of the Study:

• To comprehensively analyze cone snail venom peptides for unique Cys patterns and frameworks.
• To develop and validate a machine learning model for predicting the therapeutic potential of venom peptides.
• To assess the structural and binding similarities between novel and approved venom-derived drugs.

Main Methods:

• Analysis of 5,985 cone snail peptides across 82 species to identify Cys patterns and generate species-level pattern barcodes.
• Computation of Cys disulfide linkages from 151 Conus peptide PDB files to assess stability.
• Application of Random Forest (RF) modeling with feature extraction based on approved venom-derived drugs for therapeutic potential prediction.

Main Results:

• Generated species-level Cys pattern barcodes for 82 Conus species.
• Developed an RF model achieving 90.48% accuracy in classifying peptide therapeutic potential.
• Identified significant similarities in binding patterns between approved drugs and novel peptides predicted by the model.

Conclusions:

• Machine learning, particularly RF modeling, shows high accuracy in predicting therapeutic potential of cone snail venom peptides.
• The study highlights the utility of Cys pattern barcodes and structural analysis in drug discovery from venom.
• Further model enhancement with larger datasets and optimized feature selection can broaden its application in pharmacological research.