rtransparency

rtransparency automatically identifies and extracts indicators of research transparency from the full text of biomedical articles, in both PubMed Central (PMC) JATS XML and plain-text (PDF-derived) form. Every prediction comes with the exact statement that triggered it, so results are auditable rather than a black box. Detection is rule-based (curated regular expressions over the relevant article sections), self-contained (no GitHub-only or AGPL dependencies), and ships with reproducible accuracy benchmarks.

The ten indicators

Conflicts of interest and AI disclosure are disclosure-based: a statement on the topic counts whether the disclosure is positive or negative. Conflict-of- interest and funding statements are detected not only in English but also in Spanish, Portuguese, French, German and Italian.

Installation

# From CRAN
install.packages("rtransparency")

# Development version from GitHub
# install.packages("remotes")
remotes::install_github("choxos/rtransparency", build_vignettes = TRUE)

No GitHub-only or AGPL dependencies are required; data and code detection is native (it no longer wraps oddpub). rt_read_pdf() (PDF to text) additionally needs the poppler pdftotext utility on your system (or the optional pdftools package). The optional furrr and future packages enable parallel corpus processing, ggplot2 plotting, and jsonlite the ClinicalTrials.gov lookup.

Quick start: all ten indicators in one call

library(rtransparency)

xml <- system.file("extdata", "PMID32171256-PMC7071725.xml", package = "rtransparency")

res <- rt_all_pmc(xml)

# The predictions, one column per indicator:
res[, c("is_coi_pred", "is_fund_pred", "is_register_pred", "is_novelty_pred",
        "is_replication_pred", "is_open_data", "is_open_code", "is_ai_pred",
        "is_open_access", "is_reporting_pred")]

# Each prediction is paired with the text/value that triggered it, e.g.:
res$coi_text
res$open_data_statements
res$oa_license            # e.g. "CC-BY-4.0"
res$reporting_guideline   # e.g. "PRISMA"

rt_all_pmc() returns one row with the ten predictions, the extracted statement for each, article identifiers and metadata, the year, and is_success. is_ai_pred is NA for articles published before 2023; ai_used, ai_tools and ai_purpose say whether a disclosure reports use, of which tools, and for what. has_das records whether the article has a data-availability section.

Getting articles

rt_fetch_pmc() downloads PMC full-text XML for PMCIDs, PubMed IDs or DOIs (from NCBI by default, or from Europe PMC), reusing files already downloaded:

got <- rt_fetch_pmc(c("PMC7071725", "32171256", "10.1186/s12874-020-0914-6"),
                    dir = "xml")
res <- rt_all_pmc_dir("xml")

rt_convert_ids() maps PubMed IDs, PMCIDs and DOIs to one another. Set the ENTREZ_KEY environment variable to an NCBI API key to raise the rate limit.

Per-indicator functions

Each indicator can be run on its own, for a PMC XML file or a plain-text file:

rt_coi_pmc(xml)          # conflicts of interest
rt_fund_pmc(xml)         # funding
rt_register_pmc(xml)     # protocol registration
rt_novelty_pmc(xml)      # novelty claims
rt_replication_pmc(xml)  # replication / external validation
rt_data_code_pmc(xml)    # data AND code sharing (+ extracted links, data-availability section)
rt_ai_pmc(xml)           # generative-AI-use disclosure (2023+), with use, tools and purpose
rt_oa_pmc(xml)           # open-access status + license
rt_reporting_pmc(xml)    # reporting-guideline use + which one
rt_meta_pmc(xml)         # article metadata

Default XML namespaces are always removed, so the former remove_ns argument is no longer needed (it is accepted and ignored).

Structured metadata and follow-up checks

Some transparency signals are tagged in the JATS XML rather than written in prose, and some can be checked against outside sources:

rt_authors_pmc(xml)      # ORCID coverage of authors, CRediT contribution roles
rt_funders_pmc(xml)      # funders with Crossref Funder IDs, ROR IDs and award numbers

ids <- rt_trial_ids(res$register_text)            # NCT, ISRCTN, PROSPERO, ... numbers
rt_registration_timing(ids$trial_id)              # prospective or retrospective (ClinicalTrials.gov)

rt_fill_coi_pubmed(res)                           # COI statements recorded only in PubMed
rt_check_links(res$open_data_links)               # do the shared-data links resolve?

rt_ethics_pmc() and rt_ethics() detect ethics approval and informed consent statements. They are experimental: not yet validated against hand labels, so they are not part of rt_all_pmc(). ## Corpus-scale processing

rt_all_pmc_dir() runs all ten indicators over an entire directory (or a vector of paths). It is built for large corpora:

res <- rt_all_pmc_dir(
  "path/to/xml",          # a directory, or a character vector of file paths
  output    = "results.csv",  # resumable: re-running skips files already recorded
  parallel  = TRUE,           # via furrr + an active future::plan()
  progress  = TRUE
)

• Resumable: with output, results are written to a CSV in chunks; a re-run skips files already recorded and appends only the new ones.
• Failure-isolated: a malformed file yields an is_success = FALSE row, with the reason in error, instead of aborting the run.
• Parallel: set future::plan("multisession") and parallel = TRUE.

Plain-text input

The same detectors run on plain-text (PDF-derived) articles, given either a file path or the text itself, and return the same column names as the XML detectors:

rt_all_pdf("article.pdf")                   # all ten indicators from a PDF (needs pdftotext)
rt_all("article.txt")                       # all ten indicators from a text file
rt_all(text = rt_read_pdf("article.pdf"))   # or from text already in memory
rt_coi(text = my_text)                      # one indicator at a time

rt_all_txt_dir("path/to/txt_and_pdf")       # a whole directory, resumable and parallel

rt_ai() is the plain-text counterpart of rt_ai_pmc(). Because a text file carries no reliable publication date, it applies no 2023 year gate (it returns TRUE/FALSE, never NA) and cannot confine the scan to back-matter sections, so restrict its use to 2023-or-later articles and expect a slightly higher false-positive rate on papers that use AI as a research method.

Summarizing a corpus

Once you have one row per article, summarize the corpus:

data(rt_demo)            # a small simulated example shipped with the package

rt_summary(rt_demo)      # per-indicator prevalence with a Wilson confidence
                         # interval and a sensitivity/specificity-corrected
                         # (Rogan-Gladen) prevalence whose interval carries the
                         # uncertainty of the detector's validation

rt_summary(rt_demo, by = "year")   # subgroup summaries

rt_score(rt_demo)        # add a per-article count of openness practices met

rt_plot(rt_demo)                                  # prevalence bar chart
rt_plot(rt_demo, type = "trend", year = "year")   # prevalence over time

The accuracy correction uses the bundled rt_accuracy table (detector sensitivity and specificity for eight indicators; open-access licensing and AI-use disclosure are reported uncorrected). Supply your own estimates:

rt_accuracy                              # the bundled estimates
my_acc <- data.frame(variable = "is_open_data", sensitivity = 0.84, specificity = 0.97)
rt_summary(rt_demo, accuracy = my_acc)   # correct with your own values

Linking to FAIR assessment

The data- and code-availability links the detector extracts (open_data_links, open_code_links) can be passed to FAIR-assessment tooling such as rfair to score the findability and accessibility of the shared resources.

Validation

Every indicator is benchmarked against hand labels; the reports and the scripts that reproduce them are in inst/benchmark/ and data-raw/benchmark/. On the held-out, independently labeled test set of Serghiou et al. (2021):

Data sharing is deliberately precision-favoring, and the native data/code detector was developed against this set, so its figures are regression estimates rather than an untouched validation. The newer indicators are validated against maintainer-built, hand-labeled benchmarks:

These estimates, with the validation counts behind them, form the rt_accuracy table that rt_summary() uses to correct prevalence, and the corrected intervals carry their uncertainty. Further reports cover a five-language sample for multilingual COI and funding, plain-text parity, and Europe PMC parity. Fresh, blind validation rounds for 2025 are prepared in data-raw/validation/.

See vignette("rtransparency") for the methodology and vignette("scope-and-limitations") for what each indicator does and does not capture.

Documentation

• vignette("rtransparency"): introduction and methodology
• vignette("transparency-summary"): corpus prevalence, scoring and plotting
• vignette("ai-disclosure"): the AI-use disclosure indicator in depth
• vignette("scope-and-limitations"): indicator semantics, limitations, output schema
• Package website: https://choxos.github.io/rtransparency/

Lineage and citation

This package builds on the original rtransparent tool of Stylianos (Stelios) Serghiou, an enhanced, renamed fork maintained by Ahmad Sofi-Mahmudi (ORCID 0000-0001-6829-0823, GitHub @choxos). It adds four indicators (novelty, replication, AI disclosure, and a natively re-implemented data/code detector), multilingual COI and funding detection, plain-text parity, and corpus-scale batch processing. Serghiou is credited as an author.

The foundational paper: Serghiou et al., Assessment of transparency indicators across the biomedical literature: How open is open? PLOS Biology, 2021, doi:10.1371/journal.pbio.3001107. Run citation("rtransparency") for both references.

Use of AI

Parts of this package were developed with the assistance of generative AI (Anthropic’s Claude, via Claude Code), including code, tests, documentation, and benchmark tooling. All AI-assisted output was reviewed, run, and validated by the maintainer, who is responsible for the final content. This mirrors the kind of disclosure the package itself is built to detect.

Getting help

Please file bugs or questions as issues at https://github.com/choxos/rtransparency/issues with a minimal reproducible example.