Reflections on the journey behind the Galaxy Labs Engine paper
Dr Gareth Price reflects on the Galaxy Labs Engine paper, recently published in GigaScience, and the GCC conversations and BioHackathons that grew it from an idea into a globally deployed tool.
Dr Gareth Price, Galaxy Australia
Galaxy enables high quality reproducible research. We know this and we measure it by the number of publications that cite Galaxy. That number is huge, but a subset of that number captures publications written by members of the Galaxy community. To me, these publications have a special place as they show that we do more than just service research — we develop and innovate Galaxy’s offerings and have done so for over 20 years. Recently, our publication describing the Galaxy Labs Engine went live on GigaScience and provided me a moment to pause and reflect on this achievement.
For context, why a Labs Engine and why Labs? There is history, as I explore below, but to begin — the notion of a tailored Galaxy interface and experience is very attractive to our operation of Galaxy Australia. We have often asked ourselves why we want so many Labs? For me, it’s about giving researchers a headstart. For the individual user, we want them to easily discover the tools, workflows and training they need. Galaxy already has many great options to cater for the individual. For research communities, we want them to be able to engage with their researchers and provide the same experience we give to individuals. We also don’t want to be either the technical or domain knowledge gatekeepers to community driven development. The Galaxy Labs Engine is our solution to putting the design control of Galaxy Labs into the hands of the people who know best.
That’s the background and now I can get back to my reflections on the GLE paper. I find that we do not always “celebrate the wins” and the crafting, refining and publishing of this paper was a journey. The paper was very much a team effort and by way of apologies to my fellow authors, this blog is my rose-tinted version of events.
Many a time, often at Galaxy Community Conferences, I hear “we really should write that up” — easy to say, easy to think it’s achievable inside the conference bubble, harder to execute once back at home and committed to daily work life.
This journey to paper didn’t start with the intention to write a paper, it started with ideas and observations of our users’ use of Galaxy. It started with UI/UX! Galaxy Europe had taken the lead on providing a version of Galaxy that would speak to a subset of our users, the Galaxy flavours aka subdomains. These were a technical breakthrough, showing a version of Galaxy to someone with that interest. I want to analyse my microbiology data — there is a flavour for that! It has the tools, the workflows and tutorials that a user needs to understand this domain and the particulars for conducting high quality analytics in this space. We, in Australia, took an interest in these but like every good excuse of “when I have time”, we never quite found time. Then came our first Aussie Outpost of the Elixir BioHackathon in 2022, where we worked hard to work alongside the collective based in Europe. We took over a floor of a local hotel and worked into the wee hours of the night (biohackathon-2022). I have good memories of that time and of seeing Simon Gladman hunched over his laptop determined to “get flavours working on Galaxy Australia”, which of course he did. We could deploy subdomains, so naturally as you might have guessed our next big choice was which subdomain?!? That question was answered from two directions, global Galaxy activity and Australian research support.
At GCC2022, I found myself falling in with a crowd that, though I didn’t know it then, would become the yet-to-be-branded SPoC community. I met Wendi Bacon (Galaxy Alumni) and found that we shared a passion for deep user engagement and for providing to our users a version of Galaxy that met their specific needs. Wendi and SPoC built a Single Cell subdomain on Galaxy Europe, through iterative rounds of collaboration with the EU team. It looked good, it had the tools, the workflows and the GTN support a researcher would need to conduct their analysis. However, we just couldn’t deploy it in Australia without a complete replication, that would also be a static timestamp of the subdomain, not a living space that was needed for a fast paced community like SPoC. This was no one’s fault, it just was. By the time GCC2023 came around — and a marvelous event it was if I’m allowed a self-assessment — there was a lot of work to build the Galaxy Lab code base. Critically we ended up with Lab content being served from GitHub, realising the vision to have content become open sourced and dynamic. The next challenge was scale, how to allow more contribution to Labs and get more Labs created.
In Australia and at the time there was, and still is, lots of activity to better understand the vulnerable or threatened species of Australia. The Threatened Species Initiative were doing a lot of genome assembly and annotation tasks and wanted the capability to do these tasks accessible to a larger audience of interested researchers, ecologists, policymakers and park rangers. Accessing the BioCommons for staff skilled in UI/UX collaboration alongside domain experts, the Genome Lab was realised (galaxyau-genome-lab) using the Labs Engine.
The next timepoint that sticks in my mind is GCC2024. I have great memories of my excitement in co-planning a CoFest project with Wendi — we were going to make Labs a thing! We had Cameron Hyde (Cam) back in Australia and we had Delphine Larivière with us in Brno. We were excited. We were going to build a Lab! We actually spent most of CoFest breaking Labs. We blundered through code blocks, we tried to write in all our rule exceptions, we asked for features that weren’t even in our pipedreams — but hot damn we had fun doing it! More importantly, we got to the kernel of the problem by coming nearly full circle to a core motivation — how to take Lab generation out of the hands of SysAdmins / Web Developers and put it in the hands of biologists, researchers and community members? Or more bluntly, what could Cam develop that meant we could stop bugging him for every small code change we required?
The Galaxy Labs Engine and the wonderful support documentation at labs.usegalaxy.org.au was born. Which brought us back to the hardest observation of all — “we really should write that up”.
So, on reflection, did we end at the place we envisaged at the start? I’m not sure and that’s OK. We did learn, we built, we tested and we persisted. We got to that end. I hope in this ramble a truth of the Galaxy community comes out, that is that we do not work in isolation, we collaborate, we learn from each other, we inspire each other. Also, we have a whole heap of fun when we are together, be it GCCs, Hackathons, Galaxy Days, conferences — I love that Galaxy people love to get together and work together. This paper, this functionality in Galaxy, would not have been possible without it.
I want to thank all my co-authors for their contributions to make the Labs Engine and realise the publication. I also want to thank Nate at Galaxy Main for ever so casually dropping the Single Cell Lab onto Galaxy Main, after a conversation at (you guessed it) GCC2025. This allowed us to say with confidence that Labs could truly be deployed globally. Or, AU and EU together gave us only two sites — short of the scientific minimum of three — and with ORG we made it!
Well, actually it isn’t the end! We always saw this paper as a launching point for future Labs. So maybe I should say this is the End of the Beginning and not the Beginning of the End.
With this publication I sincerely hope we have described constructing a Lab as an easy and accessible process. We know of two Labs in active development in Australia, one for global science and one as a landing portal connecting Australian genomic services. Even now, we are not staying still — Cam is in the final stages of deploying an AI assisted Lab builder, with the aim to further reduce the technical barriers for researchers to start on the Lab journey.
The paper is available at https://doi.org/10.1093/gigascience/giag041 and the Australian BioCommons news story at https://www.biocommons.org.au/news/galaxy-labs-engine.