This week’s recap highlights an interesting new model of deep ancestral structure shared by humans unearthed using a new coalescent-based HMM (cobraa), a genomic language model for predicting enhancers and their allele-specific activity, atom-level enzyme active site scaffolding using RFdiffusion2, and a new perspective article on multimodal foundation models in biology.

Join us for an open developer forum exploring how R and Rust can work together in bioinformatics and Bioconductor development. Organised and hosted by Lluís Revilla (Bioconductor Community Advisory Board), this online session will discuss technical and practical aspects of using Rust alongside R - discussing extendr, Rust-based package development, and real-world use cases. Whether you’re Rust-curious or already experimenting, all are welcome.

Ever since Werner Heisenberg formulated the uncertainty principle, physics has been grappling with the unsettling realization that the universe does not function like clockwork—a metaphor Newton introduced and which held sway for centuries.

One more Positron-related post! It’s a quick one, just to highlight one feature I think is really neat and helpful: the Connections Pane. In a newer research project I’m working on, I have geocoded data for every foreign aid project run by most donor countries since 1989.

From April 23rd to 25th, 2025, we hosted the first Bioconductor microbiome course in Brazil, a three-afternoon workshop held at the Interunits Post-Graduate Program of Bioinformatics at the Federal University of Minas Gerais (UFMG) (link) in Belo Horizonte.

The past is powerful evidence for arguments about the present.

Warning: I do not specialize in digital literacy and my understanding of such matters is limited.

This is the sixth and the final blog post in a series on spatial machine learning with R. You can find the list of other blog posts in this series in part one.

I am not a staunch advocate of Bayesian methods — I can totally see how for some questions a frequentist approach may provide more satisfactory answers. In this post, we’ll explore how for a simple scenario (negative binomial regression with small sample size), standard frequentist methods fail at being frequentist while standard Bayesian methods provide good frequentist guarantees.

When I saw that the TidyTuesday dataset was the the XKCD color survey this week, I had to jump in! source(here::here("_defaults.R")) library(tidyverse) library(tidytuesdayR) library(tinytable) library(mgcv) library(marginaleffects) library(ggdist) library(ggdensity) library(geomtextpath) set.seed(2025-07-08) # eval: false # downloading &

In public and political discourses, 2015 was constructed as the peak of a ‘refugee crisis.’ While a crisis is supposed to be temporary by definition, it seems that many references to migration have since been associated with the label ‘crisis.’ We argue that a ‘permanentification’ of a regime of crisis has been taking place over the last years.