A MeSH-based text mining method for identifying novel prebiotics.

Area of Science:

• Microbiology
• Bioinformatics
• Nutrition Science

Background:

• Prebiotics positively influence host health by modulating gut microbiota composition.
• Identifying novel prebiotics is challenging due to stringent criteria, limiting the discovery of new candidates.
• Existing research contains valuable data on prebiotic features that can predict new candidates.

Purpose of the Study:

• To propose and validate a text mining method for discovering novel prebiotic candidates.
• To leverage structured biomedical literature (PubMed) for prebiotic identification.
• To establish an efficient approach for expanding the repertoire of known prebiotics.

Main Methods:

• Utilized a Medical Subject Headings (MeSH)-based text mining strategy on PubMed data.
• Developed an optimal feature set for prebiotic prediction using a systematic feature-ranking algorithm.
• Classified carbohydrates based on chemical and biological attributes and employed cross-validation to assess model accuracy.

Main Results:

• The developed method achieved a specificity of 0.876 and a sensitivity of 0.838 in predicting prebiotics.
• A high-confidence list of potential prebiotic candidates was identified, supported by existing literature.
• The classification algorithm effectively distinguished prebiotics from other carbohydrates based on defined features.

Conclusions:

• Text mining of large-scale biomedical literature is a viable and promising strategy for discovering new prebiotics.
• The proposed MeSH-based method offers an efficient approach to accelerate prebiotic discovery.
• This study provides a foundation for utilizing computational methods in identifying functional food ingredients.