PMC text mining subset in BioC: about three million full-text articles and growing

Area of Science:

• Biomedical Informatics
• Computational Biology
• Text Mining

Background:

• Growing interest in text mining full-text biomedical research articles.
• Need for automated processing and improved interoperability of large biomedical datasets.

Purpose of the Study:

• To convert a large corpus of full-text biomedical articles into a standardized format for enhanced text mining.
• To improve data sharing and interoperability within the research community.

Main Methods:

• Conversion of nearly 3 million full-text articles from PubMed Central Open Access and Author Manuscript subsets.
• Utilizing BioC (Bioc) as a community-driven data structure in XML or JSON format.
• Providing access to the BioC collection via File Transfer Protocol (FTP) and a Web API.

Main Results:

• A collection of millions of biomedical articles converted to the BioC format.
• Successful implementation of FTP for bulk access and a Web API for targeted data retrieval.
• Widespread adoption and use of the BioC collection by the research community since 2017.

Conclusions:

• The BioC format facilitates automated processing and interoperability of biomedical full-text articles.
• The BioC collection, accessible via FTP and API, is a valuable resource for text mining and biomedical research.
• The project successfully addresses the growing need for efficient access and analysis of biomedical literature.