MilkOligoThesaurus

MilkOligoThesaurus provides names, synonyms, and identifiers for 245 mammalian milk oligosaccharides to enable data extraction from unstructured text and support understanding of the biological significance of milk oligosaccharide variation across mammals.

Key Features:

• Terminological resources: Four terminological resources assign unique identifiers to entities, supported by external ontologies, to address data interoperability across disparate publications and databases.
• Entity coverage: The resource encompasses names, synonyms, and identifiers for 245 mammalian milk oligosaccharides.
• Semantic annotation: The annotation scheme is a rich entity/relation model designed to describe diversity patterns of milk oligosaccharides according to female factor variability and to help better understand structure-related function of milk oligosaccharides.
• Entity types: The corpus annotates entities related to individuals, samples, oligosaccharides, and oligosaccharide quantification.
• Relationship modeling: Entities are linked by binary and n-ary relationships.

Scientific Applications:

• Natural language processing training: The corpus serves as a gold standard for training and evaluating natural language processing models to extract information on milk oligosaccharides.
• Knowledge extraction: Automated natural language processing methods extract and gather structured data from unstructured texts to gain insight into the biological significance of milk oligosaccharide variation across mammals.
• Interoperability across databases: The terminological resources enable data interoperability across disparate publications and databases.

Methodology:

The corpus consists of abstracts (15) and extracts (15) from 20 full text articles indexed by PubMed, annotated according to developed annotation guidelines. Baseline information extraction models were applied to the corpus for experimental evaluation.