AI & ML Development Analytics

rOpenSci Advances Bioconductor’s R-Universe Migration and Open Science Projects

Oct 22, 2026 649 views
Sign in to save

rOpenSci’s October roundup is less about one headline than about the infrastructure and people that keep an open science community working. The clearest example is Mark Padgham, whose work has shaped both statistical software review and the tools behind it. Alongside that transition, the organization reports progress on Bioconductor’s move to R-Universe, new Champions projects, and a steady flow of packages and reviews.

You can read the full post on the rOpenSci blog. Here’s what stands out.

People and community

Thank you, Mark Padgham

Mark Padgham leaves rOpenSci at the end of October, after joining the staff in December 2019 to help extend software peer review into statistical topics. His contribution reaches well beyond that brief. He helped define how rOpenSci reviews statistical software, built much of the machinery supporting those reviews, and created R packages used across the community. If you’ve submitted, reviewed, or edited a package, there’s a good chance you’ve encountered his work.

Read more about his contributions in rOpenSci’s profile of Mark’s work. The team is also holding a farewell celebration on Wednesday, October 28 at 15:00 Central European time.

Champions at work

The 2026–2027 Champions cohort is developing projects in package development and peer review. Participants have held their first cohort meeting in two sessions to suit different time zones, sharing progress and getting feedback from fellow Champions.

Outreach is underway too. Linda Cabrera, with support from mentor Luis Verde, led a workshop on R Projects with RStudio and GitHub for the Ecuadorian Statistical Society. She also introduced attendees to rOpenSci and its programs.

The 2025–2026 cohort has now completed its program. At the closing call, Más Allá del Código: muestra abierta de los proyectos de nuestros campeon(a|e)s (Beyond the Code: an open showcase of our Champions’ Projects), participants presented work ranging from creating and reviewing R packages to community outreach. They described outcomes that included scholarships, collaborations, and professional growth. The recording and Spanish subtitles are available.

Some Champions and mentors will take that work to LatinR 2026 in Medellín, Colombia. They’ll present during the conference and join HackeaR, a hackathon focused on issues, translations, and pull requests for base R, rOpenSci packages, and its resources. If you’re nearby, you can join them.

R-Universe and Bioconductor

🧬 The Bioconductor migration to R-Universe has advanced: packages from Bioconductor’s “experiment,” “workflows,” and “books” repositories are now on bioc.r-universe.dev and bioc-release.r-universe.dev. A presentation by Sean Davis offers more detail.

Supporting those repositories required a new “data” package category in R-Universe. These packages get a larger size allowance, with a tradeoff: they’re built and checked only on Linux. That’s a practical adjustment for a substantial migration, though the update doesn’t explain how many packages will fall into this category.

🚀 Bioconductor, with about 3,000 packages, is by far the largest universe on the platform. The work has also prompted dozens of frontend and server performance improvements, intended to speed up pages and repositories across every universe. rOpenSci says a fuller account of the transition is due before year’s end.

Fifteen years of rOpenSci

The organization is continuing its series on the people behind the community. This month, Mauro Lepore recounts his experience with rOpenSci and draws lessons that extend beyond it.

A separate recap of the anniversary events covers conversations about rOpenSci’s past, present, and future with people who joined the 15-year celebration. More community stories are planned.

Upcoming coworking sessions

The coworking program has two upcoming sessions:

  • Tuesday, November 3, at 09:00 Western Australia time (01:00 UTC): “Climate Science in R”, hosted by Steffi LaZerte with co-host Elio Campitelli. The session will look at how R is used in climate research and give participants a chance to meet and talk with Campitelli.
  • Tuesday, December 8, at 14:00 Central European time (12:00 UTC): “Code Linting in R”, hosted by Steffi LaZerte with co-host Etienne Bacher. Participants can try linters on their R code and discuss linting, including flir and Jarl. December’s session is one week later than usual.

You can also cowork independently: work on an R-related task, give an overlooked package attention, or make progress on whatever is on your list.

Software additions and releases

📦 Three packages have joined the rOpenSci software suite. They cover very different work, from Brazilian public statistics to river-space mapping and evolutionary models.

  • ibger, developed by André Leite with Marcos Wasiliew, Hugo Vasconcelos, Carlos Amorim, and Diogo Bezerra, provides a tidyverse-friendly interface to the Brazilian Institute of Geography and Statistics (IBGE) aggregate data API, documented at the IBGE API documentation. It can query aggregates, variables, localities, periods, and metadata from surveys and censuses. Allan Quadros and Diana García-Cortés reviewed it through software review issue 787.
  • rcrisp, by Claudiu Forgaci and Francesco Nattino, automates the delineation of riverside urban areas using a method introduced in Forgaci (2018), described at doi:10.7480/abe.2018.31. It identifies corridor boundaries, segments the corridor, and delineates river space using street-network and digital-elevation data in a projected coordinate reference system. The results can help characterize spatial patterns around rivers. Mark Padgham and Anatolii Tsyplenkov reviewed it in issue 718.
  • coevolve, developed by Scott Claessens and Erik Ringen, fits Bayesian generalized dynamic phylogenetic models using Stan or JAX. It supports an arbitrary number of variables with different response distributions evolving together on a phylogenetic tree under an Ornstein–Uhlenbeck process, letting researchers assess directionality between variables. Its optional JAX backend supports high-performance sampling with NumPyro and potential GPU acceleration via reticulate. Rob Moss and Krzysztof Bartoszek reviewed it in issue 717.

Browse more rOpenSci packages or learn about Software Peer Review.

Sixteen packages also received updates since the previous newsletter:

Software Peer Review

Seventeen submissions are recently closed or active, with four more on hold. The active queue spans several review stages:

Software peer review: seven packages in the queue

Seven submissions are currently at the “1/editor-checks” stage of rOpenSci’s Software Peer Review process. The range is telling: the queue includes tools for marine analysis, survey data, statistical work, and multilingual scientific computing. If you’re interested in reviewing or submitting a package, rOpenSci explains how to take part in Software Peer Review.

On the blog

Three recent posts look at rOpenSci’s history and the people who have shaped it.

Ways to contribute

Two R packages need maintainers: suppdata, which retrieves supplementary data in R, and ghql, an R client for GraphQL. If you’re considering taking one on, start with What Does It Mean to Maintain a Package?; volunteering discussions are open in the suppdata issue and the ghql issue. For other contribution opportunities, check the help wanted page. Before opening a pull request, ask in the relevant issue whether the work is still needed.

Package development corner

Some useful information for R package developers. 👀

A welcoming place to learn pull requests

Jonathan Carroll wrote about a package repository designed for newcomers to practice making real pull requests. That kind of low-pressure practice space can help people cross the gap between wanting to contribute and feeling ready to do it.

“My target audience for this repository is people who wish they had somewhere to send a real pull request because they don’t yet feel comfortable participating in open source software.”

Catch tests that change global state

A test can pass while quietly leaving the environment in a different state. testthat::set_state_inspector() helps catch that: register values you expect to stay stable, such as environment variables, and the inspector checks them before and after each test. The vcr example tracks temporary files, files in the working directory, and the VCR configuration: 
testthat::set_state_inspector(\() {
temp_files <- dir(tempdir())
temp_files <- temp_files[!grepl("^callr", temp_files)]
temp_files <- temp_files[!grepl("^webfakes", temp_files)]
list(
temp_files = temp_files,
wd_files = dir(),
vcr_config = the$config
)
})
Steffi LaZerte spotted the function in Hadley Wickham’s and Jenny Bryan’s workshop materials, which Mauro Lepore had shared in rOpenSci’s Slack workspace.

Run selected tests with just

The filter argument in testthat’s test function lets you run tests whose names match a string. Etienne Bacher’s post on using just shows how to define aliases that make this targeted test workflow easier to repeat. The command-line tool has other uses too. For example, you can define just lint to run Jarl in the current directory, using just.

Git’s hash transition deserves attention

Git plans to change its content hashing algorithm. The implications aren’t fully explained by the announcement alone, but the change could affect workflows and tooling that depend on Git’s hashes. Scott Chacon, author of the Pro Git book, lays out why the shift matters and where it may cause problems in this post about Git and SHA-256.

Stay connected

If you want to take part in rOpenSci, its Contributing Guide points to code and non-code contributions, including sharing how you use the projects. You can also support the organization through donations. Sign up for the newsletter through the newsletter form. Between issues, follow rOpenSci on its website, Mastodon, or LinkedIn. Thanks for reading; see you next time.
To leave a comment for the author, follow the link and comment on their blog: rOpenSci - open tools for open science.
R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
To share your content on R-bloggers, click here if you have a blog, or here if you don't.
Continue reading: rOpenSci News Digest, October 2026
Source: rOpenSci · www.r-bloggers.com

Comments

Sign in to comment.
No comments yet. Be the first to comment.

Related Articles

rOpenSci News Digest, October 2026
TF
Tech Frontier