Using an Ensemble to Identify and Classify Macroalgae Antimicrobial Peptides.

Area of Science:

• Marine Biology
• Biotechnology
• Computational Biology

Background:

• The rise of multi-drug resistant microbes necessitates novel therapeutic strategies.
• Antimicrobial peptides (AMPs) are a promising natural alternative, exhibiting broad-spectrum activity against pathogens and cancer.
• Limited data exists on AMPs from marine macroalgae, despite advancements in prediction methods.

Purpose of the Study:

• To develop and validate a computational approach for identifying and classifying antimicrobial peptides (AMPs) from marine macroalgae.
• To explore the potential of marine macroalgae as a source of novel AMPs.

Main Methods:

• A two-tier ensemble of heterogeneous machine learning models was designed.
• The first tier used binary classifiers to predict AMPs from protein sequences.
• The second tier employed a multi-class ensemble to categorize identified AMPs by functional family.

Main Results:

• The model successfully identified 39 putative AMP sequences from 12 different macroalgae species across three phyla.
• The ensemble approach demonstrated efficacy in both AMP prediction and functional family classification.

Conclusions:

• Marine macroalgae represent a significant, yet underexplored, source of novel antimicrobial peptides.
• The developed two-tier machine learning model provides a robust framework for discovering AMPs and can be adapted for other protein types.
• This research opens new avenues for natural product discovery in the fight against antimicrobial resistance.