User-friendly extensions to MeSH

The Medical Subject Headings (MeSH) is a controled vocabulary produced by the NLM for cataloging biomedical information. The resource is structured as an ontology and is used for PubMed/MEDLINE annotation. Here we provide user-friendly datasets derived from MeSH. Currently, two record types are processed: Descriptors and Supplementary Concept Records.

Notebooks

• descriptors.ipynb — processes Descriptors (also known as Main Headings)
• supplementary-concept-records.ipynb — processes Supplementary Concept Records (SCRs)

Datasets

The data directory contains the created datasets:

• terms.tsv — table of Descriptor terms.
• descriptor-terms.tsv — table of Descriptor names.
• mesh.json — a JSON-formatted representation of the Descriptor ontology. Includes term identifiers, names, semantic types, parents, and tree numbers.
• ontology.gexf.gz — a GEXF representation of the descriptor ontology that is compatable with newtorkx.
• symptoms.tsv — symptom Descriptors (the 438 descendants of D012816
• tree-numbers.tsv — table of tree numbers for each Descriptor. A tree number represents a path to the the root. This file is handy for mapping to external resources which occasionally identify MeSH Descriptors by their tree numbers (a bad but prevalent practice).
• supplemental-records.tsv — table of SCR terms.
• supplemental-terms.tsv — table of SCR names.

License

This repository is released as CC0