Overview
dbparser is an rOpenSci peer-reviewed R package that parses and integrates major pharmacological databases into standardized, analysis-ready R objects called dvobjects (drugverse objects).
Pharmacological databases use incompatible formats and structures, forcing researchers to write custom parsing scripts — a process that consumes 60–80% of analysis time. dbparser eliminates this bottleneck with unified parsing functions, chainable merge operations, and a consistent output structure that enables reproducible, cross-database analyses.
With recent updates, dbparser has evolved into an integration engine, allowing you to merge mechanistic data (DrugBank) with real-world phenotypic data (OnSIDES) and drug-drug interaction risks (TWOSIDES).
Installation
# From CRAN (stable)
install.packages("dbparser")
# From GitHub (development)
# install.packages("pak")
pak::pak("ropensci/dbparser")Supported Databases
| Parser | Source | Reads | Tables returned | Licence of the source data | Since |
|---|---|---|---|---|---|
parseDrugBank() |
DrugBank | The full database XML, or the zip it comes in |
drugs; with options also salts, products, references, cett
|
not stated | 2.0.0 |
parseOnSIDES() |
OnSIDES | A folder with product_label.csv, product_adverse_effect.csv, product_to_rxnorm.csv, vocab_meddra_adverse_effect.csv, vocab_rxnorm_ingredient.csv, vocab_rxnorm_product.csv, vocab_rxnorm_ingredient_to_product.csv, optionally high_confidence.csv
|
One table per file: product_label, product_adverse_effect, product_to_rxnorm, vocab_meddra_adverse_effect, vocab_rxnorm_ingredient, vocab_rxnorm_product, vocab_rxnorm_ingredient_to_product, optionally high_confidence
|
CC-BY-4.0 | 2.2.1 |
parseTWOSIDES() |
TWOSIDES | TWOSIDES.csv.gz |
drug_drug_interactions |
not stated | 2.2.1 |
parseHPO() |
Human Phenotype Ontology, as distributed by Harmonizome |
gene_attribute_edges.txt, plain or gzip-compressed |
associations, genes, phenotypes
|
CC-BY-NC-SA-4.0 | 2.3.0 |
parseHGNC() |
HGNC |
hgnc_complete_set.txt, optionally withdrawn.txt
|
gene, gene_name, withdrawn
|
CC0-1.0 | 2.3.0 |
parseHealthCanadaDPD() |
Health Canada Drug Product Database |
allfiles.zip, allfiles_ap.zip, allfiles_ia.zip, allfiles_dr.zip (or their unzipped folders) and inactive.zip
|
drug_product, active_ingredients, companies, form, packaging, pharmaceutical_std, route, schedule, status, therapeutic_class, veterinary_species, biosimilars, inactive_products
|
OGL-Canada-2.0 | 2.3.0 |
parseOpenFDA() |
openFDA |
drug-drugsfda-0001-of-0001.json.zip, drug-ndc-0001-of-0001.json.zip (or the unzipped JSON) |
drugsfda_application, drugsfda_product, drugsfda_active_ingredient, drugsfda_submission, drugsfda_submission_property_type, drugsfda_application_doc, drugsfda_openfda, ndc_product, ndc_active_ingredient, ndc_route, ndc_pharm_class, ndc_packaging, ndc_openfda
|
CC0-1.0 | 2.3.0 |
Each source keeps its own licence; check its terms before use. “Not stated” means the source’s own pages give no licence for the data; every parsed object records the source’s licence, terms page and requested citation (see get_source_metadata()).
The DrugBank XML is downloaded from DrugBank with a free account; versions 5.1.0 through 5.1.12 have been tested. If you find errors with any DrugBank version, please submit an issue.
Tutorials: Getting started, and the worked examples on DrugBank, integrated pharmacovigilance and HPO target phenotypes.
Quick Start
Parse a Single Database
library(dbparser)
# Parse DrugBank
drugbank_db <- parseDrugBank("data/drugbank.xml")
# Parse OnSIDES
onsides_db <- parseOnSIDES("data/onsides/")
# Parse TWOSIDES
twosides_db <- parseTWOSIDES("data/TWOSIDES.csv.gz")Integration Pipeline
The power of dbparser lies in its ability to chain parsers and mergers together. Here is how you can build a complete pharmacovigilance dataset:
library(dbparser)
library(dplyr)
# 1. Parse the raw databases
drugbank_db <- parseDrugBank("data/drugbank.xml")
onsides_db <- parseOnSIDES("data/onsides/")
twosides_db <- parseTWOSIDES("data/TWOSIDES.csv.gz")
# 2. Build the Integrated Knowledge Graph
# DrugBank serves as the hub. Chain the merges.
final_db <- drugbank_db %>%
merge_drugbank_onsides(onsides_db) %>%
merge_drugbank_twosides(twosides_db)
# 3. Analyze Results
head(final_db$integrated_data$drug_drug_interactions)For a detailed case study, see the Integrated Pharmacovigilance Vignette.
The dvobject Structure
Every parser returns a dvobject: a named list of tibbles (DrugBank’s are nested in lists by section), with a source_metadata record of the source’s release, licence and file hashes. The tables and columns each parser returns are listed in dvobject structures by parser.
Use dbparser from Claude
dbparser.mcp lets Claude Code and Claude Desktop parse and read your local database files through dbparser, with each dataset’s source, release and licence in every answer. Install:
remotes::install_github("interstellar-egypt/dbparser.mcp")Commercial support and extensions
dbparser is free and open source, and it stays that way: every parser for an open database is part of this package.
It is maintained by Interstellar, which also offers:
- dbparser Pro: loads licensed sources (for example DrugBank, under your own licence) and your private data into PostgreSQL, matches products and ingredients across sources, with scheduled refresh, release-to-release comparison and validation reports.
- Services: data integration, custom sources, analysis apps and training for life-science teams.
- DrugVerse Scope: one sourced evidence page for any drug or target, built on dbparser.
Questions or early access: info@interstellar-egypt.com
| dbparser (free, open source) | dbparser Pro | |
|---|---|---|
| Parsers: DrugBank XML (from your own download), OnSIDES, TWOSIDES, HPO, HGNC, Health Canada DPD, openFDA (Drugs@FDA, NDC) (more coming). Each source keeps its own licence. | ✓ | ✓ |
| dvobject format with source, release, licence and checksum metadata | ✓ | ✓ |
| Merge DrugBank with OnSIDES, TWOSIDES and HPO, and subset parsed DrugBank and OnSIDES datasets | ✓ | ✓ |
| Documentation, vignettes and examples | ✓ | ✓ |
| Community help on GitHub Discussions | ✓ | ✓ |
| Cross-source matching: products, ingredients and salt forms joined across DrugBank, Health Canada, FDA and other registers | ✓ | |
| Licensed sources (for example DrugBank, under your own licence) and your private data loaded into PostgreSQL, with a licence gate per deployment | ✓ | |
| Your private data joined to public sources | ✓ | |
| Scheduled refresh when a source releases | ✓ | |
| Release-to-release comparison | ✓ | |
| Validation reports | ✓ | |
| DrugVerse Scope integration | ✓ | |
| Setup help and priority support | ✓ |
Used in published research
- Hammoud Z, et al. (2025). Multipath2.0: extending multilayer reproducible pathway models with omics data. Computer Methods and Programs in Biomedicine Update. doi:10.1016/j.cmpbup.2025.100189
- Su Y, et al. (2024). Multi-ancestry proteome-phenome-wide Mendelian randomization offers a comprehensive protein-disease atlas. medRxiv (preprint). doi:10.1101/2024.10.17.24315553
- Rischke S, et al. (2023). Machine learning identifies right index finger tenderness as key signal of DAS28-CRP based psoriatic arthritis activity. Scientific Reports. doi:10.1038/s41598-023-49574-4
- Namiot ED, et al. (2023). The international clinical trials registry platform (ICTRP): data integrity and the trends in clinical trials, diseases, and drugs. Frontiers in Pharmacology. doi:10.3389/fphar.2023.1228148
- Parolo S, et al. (2023). Single-cell-led drug repurposing for Alzheimer’s disease. Scientific Reports. doi:10.1038/s41598-023-27420-x
- Schubert NMA, et al. (2022). Transcriptome-guided identification of drugs for repurposing to treat age-related hearing loss. Biomolecules. doi:10.3390/biom12040498
- Mercatelli D, et al. (2022). Detection of pan-cancer surface protein biomarkers via a network-based approach on transcriptomics data. Briefings in Bioinformatics. doi:10.1093/bib/bbac400
- Yang C, et al. (2021). Mapping the landscape of synthetic lethal interactions in liver cancer. Theranostics. doi:10.7150/thno.63416
- Pérez-Moraga R, et al. (2021). A COVID-19 drug repurposing strategy through quantitative homological similarities using a topological data analysis-based framework. Pharmaceutics. doi:10.3390/pharmaceutics13040488
- Hammoud Z, Kramer F (2020). Mulath: an R package to generate integrated reproducible pathway models. Biology. doi:10.3390/biology9120483
dbparser is listed in the CRAN Epidemiology Task View and described in our JOSS paper.
Ecosystem
| Package | Description | Links |
|---|---|---|
| dbdataset | Pre-parsed DrugBank datasets ready for analysis | GitHub |
| covid19dbcand | COVID-19 drug candidate datasets | GitHub |
| periscope2 | Shiny framework for interactive dashboards | CRAN |
Citation
If you use dbparser in published research, please cite our JOSS paper:
Ali et al., (2026). dbparser: An R Package for Parsing and Integrating
Pharmacological Databases. Journal of Open Source Software, 11(118),
9950, https://doi.org/10.21105/joss.09950
citation("dbparser")If you find dbparser useful, consider ⭐ starring the GitHub repository and sharing it with colleagues.
Contributing
We welcome contributions! Please review our Contributing Guide.
Please note that the dbparser project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.
