pmparser

pmparser converts PubMed/MEDLINE XML data into a relational database to enable large-scale, reproducible analyses of biomedical literature.

Key Features:

• PubMed XML parsing: Downloads publicly available PubMed XML files and parses PubMed/MEDLINE records into a relational database schema.
• Automated updates: Incorporates regular updates from PubMed to keep the database current.
• NIH Open Citation Collection integration: Combines PubMed data with the NIH Open Citation Collection to enrich citation data.
• PMDB relational implementation: Provides PMDB, a relational database populated with parsed PubMed records containing over 31 million PMIDs.
• PostgreSQL backend: Implements PMDB using PostgreSQL 12 for database storage and scalability.
• Reproducibility: Stores bibliographic data in structured tables to support reproducible and transparent analyses.

Scientific Applications:

• Meta-analysis and systematic reviews: Enables large-scale retrieval and aggregation of PubMed records for systematic literature synthesis.
• Citation network and trend analysis: Supports construction and analysis of citation networks and temporal research-trend investigations using integrated citation data.
• Machine learning model development: Supplies extensive bibliographic and citation features for training and evaluating machine learning models on biomedical literature tasks.

Methodology:

Downloads PubMed XML files, parses PubMed/MEDLINE records into relational database tables, merges parsed data with the NIH Open Citation Collection, stores the PMDB implementation in PostgreSQL 12, and applies regular updates from PubMed.