Bibliographical analysis code and data for holobiont-related terms from 2000-2025

Code and data for analyzing broad bibliographic trends in articles that match holobiont-related terms among all microbiome-related articles. This analysis is an introductory result used in a forthcoming manuscript by Gavin M. Douglas and Andrew Inkpen. Through academic access, I was provided API key, which I placed in ~/.config/pybliometrics/pybliometrics.cfg locally. I had to configure this locally by running this script: config_scopus_api.py. Doing so produced a config file for the pybliometrics Python package to refer to, which was used to interact with the Scopus API below. I then used the script download_from_scopus_api.py (on April 20, 2026) to download all article metadata for articles that contained holobiont-related terms in their abstract, title, or keywords. Separately, we also identified all articles that matched any microbiome term, regardless of whether they matched a holobiont-related word too. We restricted matches to be between 2000-2025, to be in English, to be either articles or reviews, and to be published in journals. We searched for three groupings of holobiont-related terms: (1) holobiont; (2) metaorganism or meta-organism; (3) hologenome or hologenomic. These were the microbiome terms: microbiome, microbiota, "microbial community", or metagenome. The matches to holobiont-related terms did not also require matches to microbiome terminology (and vice versa). The matching files were written to scopus_downloads/holobiont.csv, scopus_downloads/metaorganism.csv, and scopus_downloads/hologenome.csv. The file containing matches to any microbiome term regardless of matches to holobiont-related terminology were written to scopus_downloads/microbiome_by_year/ for each separate year. These outputs were separated by year so as to not reach a rate limit when downloading from the Scopus API (see the Python script for details). All output files were then gzipped. Note that at this stage we did not have information on which articles (or other material) were found in multiple tables. We filtered these files using summarize_holobiont_related_scopus_matches.R, which performed these steps: - First converted document types, and DOIs and titles to lowercase only. - Ensured that all articles have DOI (or in cases where they are missing, placeholders were added -- including for two articles that were manually identified to overlap between the holobiont and hologenome papers). - Removed duplicate entries based on same DOIs and titles (with manual checking these were indeed the same articles). - Removed 11 citations where the usage of holobiont-related terminology was in a non-microbiome context or had an entirely different meaning (the titles of these removed citations are indicated in this R script). This script was also used to categorize each entry across these files as either specific to each search term (e.g., holobiont-only), or multiple (e.g., holobiont and hologenome), under the column "category". The combined table of all filtered and dereplicated entries across the tables was written to: scopus_downloads/summarized_focal_matches.tsv.gz. I then ran summarize_general_microbiome_scopus_matches.R to perform the same filtering and dereplication. Any matches to holobiont-related terms missing from this general set were also added in, to ensure this represents a superset. These missing matches correspond to cases where there is a match to a term like "holobiont" but not also to "microbiome", despite terms related to host-microbe interactions generally occurring in the article body. The processed table was written to scopus_downloads/summarized_general_microbiome_matches.tsv.gz. Finally, we created Figure 1 with the R script plot_scopus_citation_breakdown.R. These plots were displayed from 2004-2025 due to very few holobiont-related publications from 2000-2003.