This repository contains the code and the dataset used for our manuscript. Module 3.0 enables automated extraction of titles and abstracts of indicated articles by PubMed ID. Web-based application of module 3.0 (Pubmed exporter) is available here.

Software requirements

• Python >=3.9.12
• Numpy >=1.23.5
• Pandas >=1.5.2
• beautifulsoup4 >=4.12.0
• urllib3 >=1.26.15
• matplotlib >= 3.6.2
• matplotlib-venn >= 0.11.9
• pillow >=9.3.0
• wordcloud >=1.8.2.2
• spacy >=3.4.4
• scikit-learn >=1.0.2

Modules

• Module1.0 for producing Table1.
• Module2.0 for producing Table2 from Table1. Table3 (corresponding to Supplementary Table S1) was prepared by data cleaning of Table2.
• Module2.1 for producing Table4 (corresponding to Fig.2D) from Table3.
• Module2.2 for venn-diagram in Fig.2C.
• Module3.0 for retrieving PubMed records of the indicated publications.
• Module4.0 for word cloud analysis. WC1–4 (corresponding to Fig.3C) and WC5–8 (corresponding to Fig.3D) were prepared from Table11–14 and 15–18, respectively.
• Module5.0 for Latent Dirichlet allocation (LDA). Table7–10 were used for LDA, producing Table19–64. Table29 (topic = 20) was used for in Fig.4. Table41 includes perplexity of each topic (Supplementary Fig.S2).

PubMed exporter

The ready to use web-based application for obtaining records comprising abstracts and titles of indicated publications.