Our work on 'hidden diversity' in unbinned contigs is now published in @natmicrobiol.nature.com :

www.nature.com/articles/s41...

See the linked threads for more details!

We built an app to explore the data too at: urban-soil-mags.big-data-biology.org

Preprint + data are fully available. Urban microbiomes are still underexplored

www.biorxiv.org/content/10.6...

Antibiotic resistance landscape: many latent ARGs, but relatively few established ones. Suggests a large reservoir of uncharacterized resistance potential, but also nuance is needed when interpreting these results!

Also: >2 million small protein families linked to defense systems and mobile elements. These remain systematically overlooked

We identified >30,000 biosynthetic gene clusters, highly contiguous thanks to long reads

From 58 urban soil samples (Shanghai & Nanjing), deep sequencing (≥40 Gbp/sample) enabled recovery of 7,949 MAGs (incl. 1,060 near-finished genomes)

97% of species-level genome bins do not match a reference genome. Urban soil is largely uncharted microbial territory

Urban soils are a small part of the world and lack the glamour of the wilderness or the economic importance of agriculture, but as most people live in cities, they are very important

Our latest preprint using long-read metagenomics reveals massive hidden diversity and function in city soils

🧵

We put these out 4x/year (never more than that!), so if you sign up, we won't be spamming your mailbox with every little update!

We also have a lot of tool updates:

1. GSMC-mapper had a lot of polish put on with more checking and faster downloads
2. SPIREpy was updated to work with the newer SPIRE API (in anticipation of SPIRE2)
3. Jug now ships with skills to help use it with AI coding assistants

We had described collecting the data all the way back in 2023. Now, the analysis is finally out as a preprint

www.biorxiv.org/content/10.6...

BDB-Lab March 2026 Updates What lurks in urban soil microbiomes

Quarterly updates from the group!

Focus is our latest work on urban soil microbiomes.

They are far richer than might be expected. Our latest work using long-read metagenomics reveals massive hidden diversity and function in city soils

bigdatabiology.substack.com/p/bdb-lab-ma...

The original paper describing the resource is Duan et al., 2024

www.nature.com/articles/s41...

Also, the full database is available online at gmsc.big-data-biology.org

GMSC-mapper/ChangeLog at main · BigDataBiology/GMSC-mapper Contribute to BigDataBiology/GMSC-mapper development by creating an account on GitHub.

Several bug fixes: corrected contig length check, duplicate sequence filtering, and a filename typo (filterd → filtered)

Also, DIAMOND and MMseqs2 must now be pre-installed. GMSC-mapper no longer auto-downloads them.

github.com/BigDataBiolo...

What's new:

1. database downloads are now much smaller (compressed indexes decompressed on the fly)
2. outputs are version-stamped for reproducibility
3. there's a new `gmsc-mapper citation` subcommand
4. better error messages

GitHub - BigDataBiology/GMSC-mapper Contribute to BigDataBiology/GMSC-mapper development by creating an account on GitHub.

We released GMSC-mapper v0.2.0! GMSC-mapper queries the Global Microbial smORFs Catalog to find and annotate small proteins from metagenomic data

github.com/BigDataBiolo...

If you use Jug in your research, please cite:

Coelho, L.P., (2017). Jug: Software for Parallel Reproducible Computation in Python. Journal of Open Research Software. 5(1), p.30.

doi.org/10.5334/jors...

Bugfixes: support for Python 3.14, fixes for dict_store on Python 3, and a fix for task describe.

Full changelog: jug.readthedocs.io/en/latest/hi...

Also new:

Project-local config files: Jug now discovers .jugrc files by walking up the directory tree to the git root. Keep project-specific configuration alongside your code

Faster polars DataFrame saving. The file store now special-cases polars DataFrames for significantly faster serialization

New in 2.5.0: AI assistant integration. Jug now ships a built-in skill for Claude Code and Codex. Install it with one command:

jug install-skills --output ~/.claude/skills

Then ask your AI assistant to write jugfiles, debug stuck tasks, or explain dependencies

Jug 2.5.0 is out! Jug is a Python framework for parallel & reproducible computation. Write plain Python, run it across many processes or machines with no message-passing code.

pip install jug --upgrade

(Or use the conda-forge packages with conda/pixi)

I will be speaking at BrisJAMS next week!

BDB-Lab January 2026 Updates New Year!

New Year updates from the group with a focus on Anil Pokhrel who is working on metagenomics for food security!

bigdatabiology.substack.com/p/bdb-lab-ja...

It's time for the 2025 Gibbons Lab Research Roundup!

Yeehaw! 🤠 @isbscience.org

In these dark times, let the joy of scientific discovery and trainee success be your balm ❤️‍🩹

🧵...

Today I had the pleasure of meeting the 🇵🇹 Microbiome HD Lab

Great to see Bich Ngoc Do's work on the Portuguese gut resistome data using argNorm (github.com/BigDataBiolo... by @bigdatabiology.bsky.social). Impressive, methodical research!

@anasalmeida.bsky.social thank you for the invitation!

Creation of new junior research groups at the Institut Pasteur - Call for applications 2026 - Research The Institut Pasteur is launching an international call to recruit new junior research group leaders leveraging cutting-edge transdisciplinary approaches to exploring infectious diseases, host-microbe...

🔬 Call to create junior research groups at the Institut Pasteur

Focus: Infectious diseases, host-microbe interactions, vaccines
Special interest: AI methodologies

📅 Deadline: Feb 9, 2026
👥 2-12 years post-PhD

Apply now 📝 research.pasteur.fr/en/call/crea...

#JobOpportunity #Research

The manuscript is still current as to the basic approach and benchmarks, but the implementation is more up to date

peerj.com/articles/105...

Install from conda

anaconda.org/channels/bio...

or pip

pypi.org/project/macrel

GitHub - BigDataBiology/macrel: Predict AMPs in (meta)genomes and peptides Predict AMPs in (meta)genomes and peptides. Contribute to BigDataBiology/macrel development by creating an account on GitHub.

New version of macrel released (v1.6.0)

The biggest change is internal, using a much better approach to saving and loading models (thus removing the dependency on particular versions of scikit-learn)

A few other bugfixes were included

github.com/BigDataBiolo...