Missing something? View this page on the old site Exploring Microbial Dark Matter: Outcomes of the FAIRyMAGs Hackathon 2025The biological world as we know it is largely composed of dark matter — microorganisms that cannot yet be cultivated in the laboratory. Modern sequencing technologies and bioinformatics tools now allow us to explore this hidden world by generating metagenome-assembled genomes (MAGs). This remains a challenging endeavor, and every bit of collaboration helps to deepen our understanding of these uncultured organisms.From October 6–9, 2025, the FAIRyMAGs project (funded by ELIXIR) organized a hybrid, four-day hackathon as part of the ELIXIR BFSP Programme.The FAIRyMAGs project was started in January 2025 for the duration of 20 months until August 2026 and is led by Paul Zierep and Bérénice Batut, in collaboration with four ELIXIR Nodes:The FAIRyMAGs project aims to advance metagenomics research by developing, optimizing, evaluating, and disseminating robust FAIR workflows for building metagenome-assembled genomes (MAGs).The project has already published an initial Galaxy MAGs workflow and successfully applied it to both CAMI benchmarking datasets and real-world use cases, including termite, cloud, marine, and bee microbiomes.To further enhance the quality and reproducibility of MAG analyses, FAIRyMAGs is also improving and extending existing MAGs benchmarking frameworks by integrating the CAMI infrastructure into the Galaxy platform. The CAMI challenges provided simulated benchmark datasets for short and long reads, that can be used to benchmark assembly tools and binners used for MAGs workflows.An additional focus of the project is the study of computational resource requirements for MAG construction. The team has investigated the resources needed for the assembly step of the workflow using data provided the MGnify team.The hackathon aimed to build on the preliminary work of the project, but also openly invited the MAGs community to discuss and hack on any ideas related to MAGs generation !A total of 23 researchers from around the world participated — 8 gathering in Freiburg (Germany) and 6 in an Australian outpost (coordinated by Tiff Nelson), while the rest joined remotely.Preparation included a shared Google Docs, spreadsheets, and slides for coordination. During the hackathon, participants joined joint sessions between Australia and Europe in the mornings and Europe-focused working groups in the afternoons. Communication flowed via a dedicated Slack channel in the Galaxy Training Network (GTN).During the hackathon, researchers worked independently on various goals related to MAGs workflows, depending on their background and expertise. Researchers familiar with Galaxy updated and added Galaxy tools, and modified and extended the preliminary MAGs workflow. Some participants focused on individual MAGs generation projects—for example, Stefan Kranz adapted the workflow to support long-read input. Impressively, the full modification was completed in just one hour, showcasing the efficiency of the Galaxy workflow editor.Beyond hands-on development, the hackathon also included multiple discussion sessions that extended beyond Galaxy-centric topics. One such discussion focused on the potential to predict the computational resources required for MAGs workflow tools based on input data characteristics and metadata. Reducing resource demands could have significant environmental benefits, given the scale of MAGs-related analyses.Another collaborative effort among MAGs developers from different communities centered on establishing shared benchmark and CI-testing datasets. As part of this, core developers from the nf-core, MGnify, and Galaxy MAGs workflows initiated the first steps toward a benchmark focusing on complete MAGs workflows.The team also began developing detailed training materials on MAGs generation. The broader community was invited to contribute their own resources via the Galaxy Training Network (GTN), which supports training materials both within and beyond the Galaxy ecosystem. For example, MGnify plans to develop training modules on MAGs submission using various approaches—such as command-line, front-end submission, and Galaxy-integrated tools.All progress was tracked in the 👉 FAIRyMAGs Hackathon – Coordination & Tracking Sheet. Some of the highlight outcomes are summarized here:Plots modularized for reusabilityDiscussions on multi-sample binning based on Han et al, Nature Communications, 2025MAGs submission workflow: WIP in collaboration with the MGnify team:Updated tutorials to include missing tools:New tutorial:Using CAMI infrastructure and real datasets, the group worked to standardize benchmarking for MAG workflows.MGnify, nf-core/mag, and Galaxy teams aligned on common benchmarking datasets.Started discussion with the MAGNETO developers for benchmark alignment.Agreed datasets:Improved visualizations for CAMI II benchmarks based on amber - a dedicated MAGs benchmarking tool, originally developed for the CAMI challanges → Issue #66Added amber CAMI workflow to IWC → PR #924MAGs of the marine use case: Completeness, Contamination and Average Gene Length.Correlation between sample metadata and memory usageA key challenge identified:Galaxy cannot yet use workflow-generated parameters for dynamic resource assignment. This limits automatic resource prediction, but workarounds and future solutions are under discussion with Galaxy core developers.The FAIRyMAGs hackathon brought together a great community of MAGs scientists. And although the hackathon was slightly Galaxy-centric, a nice exchange beyond workflow engines could be established, targeting major MAGs issues together! FAIRyMAGs Hackers To continue these exchanges, we created a dedicated community Slack channel for MAGs workflows, which will be used to continue the discussions, and openly invites anyone working on MAGs. The Slack channel is part of the microbioinfo workspace. If you want to join, contact us and we will send an invite!ELIXIR Europeelixir-europe.org Open source platform for accessible, reproducible, and transparent data analysis. © 2026 Galaxy Project. All rights reserved.