A Novel Peptides Database Approach for Enhanced Dereplication of Peptaibols Using Molecular Network Based on the...

Area of Science:

• Natural Product Chemistry
• Computational Chemistry
• Mass Spectrometry

Background:

• Peptaibols are peptides rich in nonproteinogenic amino acids, produced by Trichoderma species.
• Their structural diversity and limited spectral data challenge identification using LC-HRMS/MS.
• Dereplication of peptaibols is crucial for discovering bioactive molecules.

Purpose of the Study:

• To develop a novel strategy for improved peptaibol dereplication and classification.
• To integrate theoretical MS/MS fragmentation prediction with experimental data.
• To enhance the analysis of complex peptide mixtures in high-throughput workflows.

Main Methods:

• Generation of a custom spectral database using predicted b- and y-type ions.
• Intensity modeling calibrated with experimental data.
• Dimensionality reduction using t-distributed stochastic neighbor embedding (t-SNE).

Main Results:

• The novel method demonstrated robustness across diverse peptaibols, improving sequence assignments.
• A custom spectral database guided by t-SNE facilitated fragment-based matching.
• The approach enabled fast and reliable classification of unknown peptaibols.

Conclusions:

• This study presents the first t-SNE-guided theoretical spectral database for natural product dereplication.
• The method offers a powerful tool for prioritizing bioactive peptide families.
• Further database enrichment will expand specificity and applicability to broader peptide families.