Full transcript

Podcast Ep 52: BioCommons

Welcome

Phil Ewels: Hello and welcome to the Nextflow podcast by Seqera. This is episode 52, and today we’re talking all about the Australian BioCommons Project. I’m really happy today to have Ziad and Steven on the call with me. thanks chaps, for joining. It’s a pleasure. And, we’re gonna hear all about how Nextflow and Seqera Platform are being used in Australia.

Both of you have been in the community for quite a long time. You’ve both spoken at Nextflow Summits in the past and things like that, but some people listening might not be familiar with you. Let’s just start with a little introduction.

What’s your background? How did you end up where you are today?

Ziad - introduction

Ziad Al Bkhetan: Well, first of all, thanks so much Phil for hosting us. I’m Ziad Al Bkhetan with the Australian BioCommons for more than three years. My background is in computer science and bioinformatics with more focus on bioinformatics applications in the past seven years.

I’m personally very interested in software development and developing tools and methods I make them available for researchers. That’s including what I’ve done during my PhD the University of Melbourne.

I have done some consulting roles in different companies and organizations here in Australia, Canada and the US. At Australian BioCommons, I’m a product manager and I look after some of our bioinformatics services, including the Nextflow Seqera service where we have the Platform. I make it available for Australian researchers, and also providing Australian researchers access to tier one HPC facilities at NCI and Pawsey.

Phil Ewels: Really cool.

Steven - introduction

Phil Ewels: And and Steven, how about you?

Steven Manos: I studied physics and maths at the University of Sydney and did my PhD in physics actually in optical fibers. Ended up moving to London for four years and doing a postdoc there for four years at UCL.

I guess that was my first exposure to a lot of bioinformatics, molecular life sciences and imaging work, because I was working in a lab which was engaged in a program called the Virtual Physiological Human program, the VPH program, which was part of the EU framework Initiative.

So I got a lot of exposure to, to this area through, through that project. And I think really I, I, I wanted to pursue a career in research, but ended up finding I was kind of much better at supporting other people through software engineering and, accessing compute and building online platforms for people to run analysis and so on.

And ended up, moving back to Melbourne after UCL, and coming to work as the director of research computing at the University of Melbourne.

So we grew a big team, at Melbourne doing data management and HPC and cloud computing and so on. I was responsible for building the first Nectar Research Cloud with a, a really great team that I assembled, which is an OpenStack based, research cloud that’s still running across Australia to this day.

And was there for about six or seven years, built that team. it reached a, a natural conclusion and then, had a great group of people running that place.

And then moved over to the Australian BioCommons, which is hosted at the University of Melbourne, and really took, a leadership position there looking after infrastructure. I’m essentially the lead of digital platforms and computing. So I look after cloud, HPC , our data platforms, and I look after a few actual platforms like our sensitive data platform. And we also have a bunch of other directors that are responsible for other areas, including our bioinformatics platforms, training, comms, and so on.

Phil Ewels: Fantastic. I’ll have to get you writing some Nextflow pipelines for optical physics, branch out into that field.

Introduction to Australian BioCommons

Phil Ewels: tell us a little bit about BioCommons. I mean, what is this project and hasn’t been around for that long, right?

Steven Manos: Yeah, it’s kind of really was came out, its genesis was in 2018. It was a very, very small team. Just a few people, a couple of people at that point. It was a national infrastructure investment into building national bioinformatics platforms.

We really build, national infrastructure, to support researchers across the country doing molecular, life sciences. We build and provide, digital platforms, digital software tools. We import software makes sure software’s available on various national platforms. We have a big community engagement program. We have a big training program. We operate a bunch of platforms including Seqera platform. We manage various communities, including the National Nextflow community.

And we, we have a wide range of collaborators, both people that we collaborate with, around operating these platforms, but also research collaborators and institutional collaborators that have their own unique challenges in this molecular life sciences space that we work together closely with.

I think, and this will kind of go into our why Nextflow and, and all of that, but, but really I think that the powerful thing that’s different about the BioCommons that we really have a quite a community focus. So we don’t just look at molecular life sciences as a singular blob. We actually appreciate that are within that fiefdoms that have their own needs, whether it’s genome assembly or microbiome analyses, or metabolomics, proteomics, human genomics and sensitive data, population genomics, comparative genomics.

These are all areas that are unique. They have their own needs, their own pipelines, their own workflows, and we actually build and represent and try to understand the needs of, of all those communities through the BioCommons program.

Phil Ewels: I like that. I like the term fiefdoms. In nf-core we spent ages trying to come up with what to call, what are now special interest groups. Now I’m thinking maybe is it too late? Can we go back and call them fiefdoms?

So is BioCommons its own entity or is it a collaboration across universities or what’s the setup and the funding infrastructure?

Steven Manos: That’s a, that’s a really good question. It’s, we are really part of a national program called NCRIS, which is the National Collaborative Research Infrastructure Scheme. Our parent organization is BioPlatforms Australia, and they have been around for a long time. And they, fund at various institutions, and other organizations within Australia, access to sequencing machines, metabolomics analysis and so on.

We really came about because, we were initiated really as the bioinformatics arm of BioPlatforms Australia because that was a big missing piece, of their infrastructure and obviously a valuable addition to that.

So, so we are really funded through them and in turn funded through NCRIS. BioCommons is formally homed at the University of Melbourne, there’s many other organizations we partner with. Some staff are located at the University of Melbourne, but we have other staff at QCF, the University of Queensland, the University of Sydney at the University of South Wales. we have very close collaborations with NCI and Palsy, our two tier one super computing infrastructure. So we are very dispersed nationally, which represents the dispersal of where people do research. But formally, legally speaking, it’s the University of Melbourne, which is the, the lead of the project.

And, and you ask the question of what are we, again, we kind of look like our own thing, our own company organization. But legally speaking, we are a project at the University of Melbourne in the faculty of medicine.

Phil Ewels: Okay. Really cool. And do you do any kind of direct analysis bioinformatics work, or is, is it purely helping researchers do it themselves by tooling and training?

Steven Manos: Look really our, really our focus is providing the tools for people to do that analysis. We’re not a, full stack service provider in terms of doing the bioinformatics analysis for researchers.

Generally speaking, it’s really providing training, the platforms, the tools, access to compute and so on. So we can put it in the hands of, researchers and, and not just researchers who are expert at coding, but appreciating that there’s a diversity of skills of, of high-end bioinformaticians down to, wet lab biologists who have very little, compute experience and really building tool chains and platforms that that whole spectrum can effectively use to do what’s ultimately very sophisticated, analysis of molecular data.

Nextflow usage in Australia

Phil Ewels: Where does Nextflow usage come into this picture? What’s the history of Nextflow usage within Australia?

Ziad Al Bkhetan: Sure. Well, as Steve mentioned, in our work at and BioCommons, we engage with lots of different communities and , we support them in their work.

One of the main communities we work with is the bioinformatics workflow community. And, through our engagement with them, we try to assess their interest in workflow management systems. We realized that Nextflow is one of the most popular workflow management system they use, through our engagement with us.

Some of them, they may be starting their journey with writing workflow, so they usually pick Nextflow. But also converting currently existing workflow that they are using into Nextflow for better utilization.

We also realized through the workshops we do that Nextflow workshops are one of the popular, workshops that we get a lot of requests for.

BioCommons collaboration with Seqera

Phil Ewels: Steven, maybe you can tell us a little bit about how the collaboration with Seqera started and what we’re doing now.

Steven Manos: Yeah, so, so like, many things in the BioCommons, it often involves many different parts of the BioCommons. It’s not just myself or Ziad. We have an excellent comms team and an engagement team and so on that we work with. So, Ziad was just talking about that survey and really going out and building, a workflow community and really speaking to the community what are your interests in workflows, what direction should we head in?

And that’s, I guess, an example of the fiefdoms I was talking about earlier, of trying to identify that singular interest group or that special interest group and really speak to them about what they needed.

So, really, Nextflow emerged as this leader in our focus, we had Evan, give a, talk, online. And we, we run a range of webinars, which you can find on our website on various topics. And Evan gave a very excellent talk. It was very late for him, I remember in an evening, possibly midnight his time and daytime for us as it happens with of our, timing, and time zone challenges. He gave an excellent talk just to really, start to build interest in Nextflow in Oz.

Through that survey process, we also identified the need to actually build a platform that could support users in, deploying those workflows, making ‘em available on national compute. Giving people easier access to running those workflows, and so on and so on. We also had, Evan and a bunch of others come out from Seqera and spend some time in Melbourne, and other places and giving some local talks around Nextflow and really it just kind of all went from there leading up to that real launch of our Platform when we spoke about it and announced it, at the Seqera Summit in Barcelona. In October 23.

Phil Ewels: Yeah, I remember your talk well. I was skipping back through it, this morning actually before we came on the call. It’s really fun hearing you talk about the setup.

That was quite future looking. I think you said you had about 40 or so different kind of beta testing users at the time and you were like about to go big. So. I mean, how, how did that go and what does the Seqera Platform setup look like, for BioCommons and how do you use it?

Ziad Al Bkhetan: So we currently have our platform hosted in on AWS and we support Australian researchers from different institutes, I believe by today we have more than 90 different users. We have more than 18 different research groups from different universities and research organizations accessing the platform are working on it.

We have two big institutes. We have the Garvin. They have like, we have maybe more than 10 different groups, like heavily using the service.

These users are accessing, you know, the platform some extent, a daily basis. They are at different stages. Some of them still in ex exploration phase, some of them actually using it for, for production analysis.

We also have large scale communities and consortia. We have the Australian Amphibians and reptile, initiatives that, you know, they’re part of the organization. Researchers from different universities, working specific species. We supported them with some workflows and we had them on the platform that they run regularly.

We also now working with Australian structural biology computing community as well, trying to host the most popular tools and workflows around protein folding problem on, on the platform.

Compute infrastructures in Australia

Phil Ewels: So the, you are using an enterprise installation of Seqera Platform, which is hosted on AWS as I understand. Is everyone running their compute on AWS as well, or are you supporting multiple compute infrastructures?

Steven Manos: We’ve really focused on engaging with our two tier one national supercomputing platforms, being the Pawsey Supercomputing Center that’s out in Western Australia. and then also NCI, which is the, the Supercomputer Center in Canberra.

We have done various tests on AWS and running the compute on AWS, but Australia has actually made a massive investment in tier one compute and making that available to, to our researchers and, and we’ve done a lot of work and, you know, really making those facilities, life sciences friendly and collaborating with them to make those facilities more life sciences friendly.

And part of that is making sure that external platforms actually integrate with these facilities, nicely as well. So as you can imagine, a lot of HPC is very traditionally command line based, that has a very structured access model. And really trying to take that, model and evolve it to say, well, now we have a web platform and people can, start up, and you know, type in parameters and receive their results all via our web platform, while it’s running nicely and invisibly in the backend on, on a, you know, very large computing cluster is is where we wanna get to and where what we’ve actually done.

I think that the message that we tried to get across in ‘23 as well was that we, don’t just treat the Seqera web platform as an isolated thing. We actually see it as part of a Nextflow web of infrastructure where we actually say, well, the, the platform is not just that front AWS hosted thing, it’s, it’s the platform plus the compute that sits behind it.

So a lot of our efforts today are around how do we just make that kind of web more cohesive work together? Because from a researcher’s perspective, just, you know, can I log into the Seqera platform is useful, but it’s actually not the full picture. Can I log in, select a pipeline, put in my parameters successfully, run it and get my results out? That whole end to end is what matters. So actually treating it in the end to end is where our focus is.

Phil Ewels: Absolutely makes a lot sense. I mean, there’s two different things we’re talking about really. You’ve got a Nextflow community in Australia and you’ve also got Seqera Platform, and those two might overlap, but there’s gonna be a lot of people using Nextflow who are not using Platform because they’re more comfortable using Nextflow directly and, and you’re supporting that user group as well through training and support and things as well?

Nextflow support & training

Ziad Al Bkhetan: We do support users in different context. Training and support solutions we have, you know, will target the users regardless if they use our services or not. It’s open for all Australian researchers. That’s one aspect we deal with. We support by providing access to infrastructure that will allow you to run, workflows in tier one HPC facilities that we fully subsidize, including utilizing this infrastructure through the platform.

But even just directly accessing this infrastructure, logging in and submit its large scale workflows directly through the command line as well.

We work with some of the partners, to support them to convert their workflows into Nextflow. We have done some work with the CCI Institute, the Children Cancer Institute at Sydney. They had their own, production workflows that were written in, common workflow language, CWL. And were in the, in the phase of transition into Nextflow. And we work with them closely to compare these workflows into Nextflow. Try to go from the raw reads to having fully annotated SNVs. Try to make these workflows very efficient in terms of passing the input data in addition to the parallelization you get natively from Nextflow. Users can add any resources for our mutations and get the results they they want.

So there are different aspects we support the users. Yeah. Especially when the training aspects as well. Late last year, we have done a workshop with some support from Seqera, we had Chris Hakkaart joining us here in Australia remotely. And we gave a workshop that was organized by our comms and training team with some support as well from SIH. These workshops are open for any researchers. Regardless they are using the other services we have or not, it’s open for, for, for anyone..

Highlights and lowlights

Phil Ewels: Really cool. We’re a year and a half in now from your talk in Barcelona, Steven, is there any highlights or lowlights anything that, has been more challenging than you expected or, or has worked particularly well? I.

Steven Manos: It’s a really good question. You know, the. The highlights have been that for the people that really use it, it’s amazing. And they get a lot of benefit from it. I think the other highlight is seeing the the confluence of nextflow as a workflow language, that community coalesce with, our training efforts, our community building efforts, and then the platform as well.

And of course, that’s just not an Australian thing that is very much connected to, initiatives around the world. Being able to bring all those together has also been awesome.

like the thing of like, Once you all start speaking the same language, communicating and, you know, we’re all working in Nextflow and the collaborations, and the developments that happen are actually awesome. So that’s been a really good aspect.

Think a challenging aspect is, is really, it has been the uptake of the national service, I would say has been slow, but I think it’s also maybe just the nature of the platform, but also the language as well, is that it really is about productionized workflows, about people things into the Nextflow language, porting them over, and they really need to be things that you are gonna run repeatedly, many times over, over a long period of time because there is a setup, time cost, and effort to making all this stuff, bulletproof, making it work on the platform, making it work on these backend HPCs. and that naturally I think leads you to a smaller end of end users. For those end users, and we spoke about the Garvin Institute as well as the, as well as WEHI. for them it’s, it’s been extremely beneficial and we’ll continue to do that.

Phil Ewels: Really good. Yeah. That’s something that’s always at the forefront of my mind. I think the large scale production usage of and value of Nextflow is well demonstrated and proven. but then, last year in Barcelona, Rob Syme gave us great talk about Big Nextflow and Small Nextflow. When you pass a certain threshold, and I I found the same thing, you find yourself writing Nextflow workflows for everything. Once you get familiar enough with the language that the, the cost is not very high, then it starts being useful to write everything as Nextflow. But, but there’s definitely that learning curve, like you say and it has to be worth it at every stage of your cycle.

Steven Manos: I will have to look up that talk. That sounds good.

Protein fold

Phil Ewels: Ziad Barcelona, you gave a talk in 2024. You talked, about protein fold and it was fantastic talk. Lots of really gorgeous visuals and everything. That’s clearly a real nice success story there where you’ve helped develop this pipeline.

Ziad Al Bkhetan: Protein fold project and workflow is maybe one of the great stories we have, the complexity of the problem, but also a good example actually to how Australian BioCommons operate and work in general.

As Steven mentioned, you know, we work with different life science communities. We engage with them, understand their needs. And one of the communities we worked with the Australian Structural Biology Computing in Australia. And we started this engagement with them. we talked with the champion of the community, Dr. Kate Michie at the University of New South Wales.

She was providing us with all the needs and interest for the users and researchers from this community. And it was clear that there is a need to establish a platform to support them you know, utilizing the tools in the, in this field.

One of the things that came at that time is actually that there is a protein fold workflow available in nf-core. And we thought, well, instead of just starting from scratch with the development, maybe we contribute to this workflow that’s already existed and make this benefits available for all researchers.

And the nice thing is actually the smooth process we had, so we reached out to , maintainers, the guys at the Notre Dame lab. We talked directly with Jose, one of the main maintainers and Cedric, the lab leader.

You know, there was a kind of hesitation at the beginning ‘cause you always work with the collaboration that, you know, some extra time and cost will come with the development. the way things happen actually was very smooth. They were very welcoming for the new ideas. We had some plans, they were very accommodating for the time difference. We discussed some ideas and we start to work together.

This actually started like, small scale, but just quickly actually expanded and we had more collaboration. So we have people from UNSW contributing significantly to this workflow. We have also, some support from, Sydney Informatics Hub, from us, the bio cloud team and platform team directly from Australian BioCommons. We also had PAOs lab as well working on different, models.

And now actually we are going to do another release hopefully very soon. We started with Alpha Fold 2. CollabFold, ESMfold. Now we added boltz, helix fold 3, Rosetta fold, all atoms.

And, we also added like a nice visualization component to show all the results and, you know, make this comparison just visually smoothly, even if you run the workflow direct command line, or even if you utilize it as Seqera platform, where all these services will just come out and make it easier for users to see actually the outputs of their results.

AusARG Community

Ziad Al Bkhetan: The other success story other than the protein fold was our work and support for AusARG community. This community they used to have a comprehensive workflow to do phylogenetic trees from, from raw Fastq data.

And, you know, we started this work, they reaching out to us, they had actually a bunch of Python scripts. They work on one specific machine. There is only in one guy they can reach out to or we’ll just do the analysis that other guys not there. They will have some, some, some problems. So we decided to work together and just convert the whole workflow, like make it easier to use and, and portable. And we, we did it with Nextflow.

And I work with this group like closely with Ian Brennan, who is the main developer of this workflow. It was a nice collaboration. ‘cause I’m not expert in the problem itself. I have the expertise in the development aspect. I cannot talk about the phylogenetic trees, not my, my, my area of expertise. We just complemented each other. You know, whenever I stuck, I ask. Sometimes I give them information without actually understanding everything behind. But I get, you know, like enough information to continue the development.

And from my part, I utilize my expertise in the development. And then we have this nice workflow. We deployed it in the platform and now they use it regularly. Any person actually from the team can now go to the user interface, provide a link to the input data, and they run it.

The workflow can start from different sections. Finish at any section you want. Some customization in the tools you want very smoothly, which, which I believe facilitated their work significantly.

Phil Ewels: I love these stories where, it is the same in nf-core where you have different groups of people with different expertises who can kind of come together. And when you get that mixing pot, you get some really spectacular results sometimes. Domain experts coming together with infrastructure experts and so on. Really rewarding. It’s,

Automation with Seqera Platform

Steven Manos: I guess one of the other use cases or stories that’s emerging here is, you know, following on from that structural biology story is a lot of the focus so far with Seqera platform has been that is the platform which people access natively and upfront to run protein fold or whatever pipeline.

But really it is quite a, it is, it can be API driven, it can act itself as a bit of a backend infrastructure. And, and one of our big National computational structural biology user stories that we’re responding to is, is how do we build a, a nice GUI front end, which, which gives medical researchers or other not computationally savvy people access to the sophisticated protein design tools in the backend and we’ve decided to use the Seqera platform as that go-between because it manages all those pipeline. It can manage, the compute infrastructure on the backend and then coming and building a nice, GUI on the front end. so that’s something where we’re planning on or we are working on at the moment and planning on releasing in the next six to 12 months.

Phil Ewels: A little pet project of mine is always thinking about automation. You can encapsulate a lot of complexity within platform and, you know, set it up one time.

Steven Manos: Yeah.

Phil Ewels: it makes it much easier to build automation.

Steven Manos: Yes.

Ziad Al Bkhetan: It did actually, one of the very exciting, or I would say, interesting use cases we had with automation is some work we’ve done with AGRF where they have their local HPC, but at some part of their workflows, like some jobs, they need a lot of compute, that HPC does not work. Like, you know, it’s not enough for them.

And we managed to do through the platform actually to have the configuration in a way that, you know, everything runs, runs on the HPC, but for these specific jobs they would just go to AWS directly and Nextflow will just implicitly stage the data, put them up to AWS. the whole execution, then take the data down or even play, you know, with then put an output like, you know, either locally on HPC or having them on AWS.

That was a very good use case where, situation where you might be thinking, I either have to run here or there. You could actually have this hybrid runs where some jobs goes to different infrastructure than others.

Phil Ewels: Very nice. Yeah. Nextflow abstracts away all the complexity.

Steven Manos: Yeah.

Nextflow Communty in Australia

Phil Ewels: So let’s, let’s talk a little bit about the community side of things. From my personal point of view, I’m mostly involved with nf-core, and the nf-core community started off in Europe, has had a historically quite strong European bias, has gradually spread across to the US and, and the Americas generally.

And then I feel like the last few years we’ve seen a real uptick in the number of contributions coming from Australian groups. There’s the protein fold collaboration obviously is, has been really celebrated. There’s also like the onco analyzer, pipeline which been developed by folks at the Hartwig and in Melbourne. You can start to see more and more activity and, and contributions coming from Australia or Australasia.

And then I guess the pinnacle for me really is the recent March, 2025 hackathon where it was the first time we’ve had local groups join from Australia and New Zealand.

How have you found that, kind of integration of your Australian BioCommons community with the wider Nextflow and nf-core ecosystems and communities?

Steven Manos: Yeah, so it’s a, it’s a really good question, Phil. I, I would say that the shape of the overall community, comeback to the word fiefdoms. There are many fiefdoms even within the Nextflow community, right? So, we, know, you have this nf-core effort, you have people who simply use Nextflow, you have Seqera users, we run training, you have your ambassador program and we have a, a bunch of, I mean, how many ambassadors do we have in Australia now? He had five or six or so. so, so there’s actually kind of many, many facets to this, to this Nextflow community.

And I, and I would say that the work to bring all those different things together, even around the platform itself, is an ongoing that we’re thinking about, that we’re working on, that we’re bringing together. So, so it’s, you know, and then we talk about our, our various developments as well in protein fold and other collaborations.

So, so there’s lots of kind of elements to it, which is super exciting, but it’s also complicated. and it’s something that we need to keep on top of. And even internally within the BioCommons’s we’re constantly thinking, okay, how can we do better, at engaging with this international community, but also engaging within Australia, within these various, aspects of what you call the Nextflow ecosystem.

Ziad Al Bkhetan: And I add to this one, as, as you mentioned, Phil, there is, there is I think a growth in the engagement from the Australian community or the Australian researchers, you know, the international community with time.

I think the main challenges are around, the time zone, and the geographical distance that makes it very difficult for them to participate. You know, either on online meetings or even, in person events and, and activities.

As Steven mentioned, we have now five different, Nextflow ambassadors. And they are actually, the good thing is they are in different states that hopefully they will poster these communities in these different states which will encourage other actually as well to participate in, in different workflows.

I start to see more involvement in the development aspect and some of the Nextflow workflows at nf-core, repository. We had Australia on BioCommons you know, we also trying to support all these activities, including the training activities we do, the workshops. We also, engage in different webinars and, and the presentation, talk about Nextflow about the services we have and, you know, show the users and the researchers that, you know, we can help them and support them, which hopefully will bring them together in some point.

2025 nf-core hackathon

Ziad Al Bkhetan: The hackathon as well was actually our first official participation in nf-core hackathon and it was a great experience. our comms engagement team worked with actually the researchers and we managed to organize the local side in Sydney and try to get, you know, at least the super users we have now.

We tried to have kind of a diverse from different states. So hopefully this will help to establish, you know, new local sites in different states, which will make it easier for everyone as well to participate. You know, we talk about travel distance between Australia, for example, Europe, but even within Australia. It’s not that easy as well to get everyone in, in, in one location.

Hopefully these activities will participate in improving and growing the community. And, hopefully in the next hackathon we’ll be having different local sites in different states. And we’ll be advertising all this information in our websites. So for our researchers to join.

also be doing different workshops in Nextflow We actually we have planned, a workshop at some point in June or July just for beginner users, but also doing something more advanced or running Nextflow and h HPCs specifically with examples on our tier one HPC facilities. But, you know, the same learnings can be transferred to different infrastructures for the users to utilize.

Phil Ewels: Brilliant. Yeah, I, I mean, for Nextflow community, since Day Zero has always been very kind of grassroots, bottom up kind of approach, and I think it is really exciting to see Australia and New Zealand’s start to hit this critical mass. Where it’s self-sufficient.

In the early hackathons, we did have some researchers and some users joining from Australia, but it was a bit disheartening for them, I think. ‘Cause they would log in and they’d be like two people and not really very much happening. And so I think people don’t stick around for very long if it, if there’s not really anything happening.

But by, harmonizing our community efforts. And it was a pleasure working with your comms team it’s very slick operation. But, but by having that node in Sydney, then you have a natural center of gravity. And it means there is a momentum continuing there for people who might want to come and join. And it starts to snowball after a certain point.

One of my highlights from the hackathon was waking up at six in the morning and seeing photos and all the activity that had been going on while I was asleep. And it really was like a continuous 24 hour cycle of, of as the, as the world spun, people were collaborating across the planet. It’s really, really satisfying.

Steven Manos: very cool.

Future plans for Nextflow & Seqera Platform

Phil Ewels: How do you see things playing out over the next couple of years in, in Australia with Nextflow and with Seqera Platform?

Steven Manos: I think, I think for, my perspective as more the infrastructure platform, person, I think it’s really, I mean it’s obviously growing the user base of, of our Seqera Platform, and increasing its reach, really making sure it, it is tightened in terms of its integration with our national ecosystem of compute.

We’ve seen some amazing exemplars of places like WEHI that actually use it to manage access to their own HPC infrastructure, and really being able to promote it as a platform which can be used in that way.

Yeah, really. And, and diversifying those stories. How can this platform be used to improve your organizational bioinformatics agenda? I think is, is something we just need to do more of, telling those stories.

So yeah, really growing that user base, growing our integrations, but also, ultimately, like I said earlier, making these kind of sophisticated analysis methods more widely available to a, a wide range of skill sets, within various, parts of the, the life sciences ecosystem.

Tips for building a National Nextflow platform

Phil Ewels: Having the experience that you do of setting up both Nextflow support as a service and also as Seqera Platform as a service to national group of researchers. If there’s anyone listening in other countries with similar setups who might be interested in this kind of thing. Do you have any advice? Is there anything you’d do differently or any kind of tips that you can give listeners?

Ziad Al Bkhetan: It’s a difficult question. we do, actually, one of the main challenges is, you know, establishing a national service comes with a lot of diversity on the users, their expertise, interest, their applications, their research areas. And having one platform that will be suiting everyone is a bit tricky.

We try to make things very general, very easy to use. We try to provide a lot of documentations that we think actually was very useful for our users. You know, in addition to all the documentations and online materials that was set, Seqera Platform has and Nextflow as well. We do have our own version of this documentation, just to make it easier for Australian researchers, make it, make the documentation more specific to the platform, more specific to the infrastructure we have that most likely the users will be using, providing more example. It did make it easier for everyone. It’s still a bit of a challenging issue.

Also if anyone has similar problem. We, we very happy to have a chat with them, try to discuss with them, share our experience with them, and see if there is any asset we can help and support with as well, even at international scale.

Steven Manos: When we started the the process of building this national platform, I guess we have a, checklist or a process we run through of all the things we need to consider, Ziad mentioned a bunch of them around documentation and, and what communities and what users and so on.

I think possibly one of the most valuable things we did early on was we, we only, we ran it as a beta service for a while, but we only engaged with groups who had a bioinformatician who either knew Nextflow or was willing to use or, or, or learn Nextflow.

I’m glad we did that because we would’ve otherwise been in a situation where we were doing all the development effort and the work, but it’s really like we will help you up your skills in how to use Nextflow, how to port your pipelines to the Nextflow how to run them on the then Tower platform and down the Seqera platform.

So, so actually kind of making it clear that we both have our roles and responsibilities and your responsibility is to build within your own organization that Nextflow expertise.

I think it was really good because we had a, we had a mixed bag of, of people that did have that expertise and they were ready to go, but others that, that didn’t initially have that.

We had one, children’s institute that didn’t initially, but then they came back about a year or two later and said, we’ve got funding, we’ve hired a bioinformatician and we’re ready to go. And they’ve been doing amazing things ever since.

So that was kind of a very deliberate decision to try and build local Nextflow expertise and, and home it in these different places rather than just people expecting or seeing it be available elsewhere.

So I think that was quite, I. Quite a valuable thing and it, it’s a big, big tip I would provide especially if you’re gonna build a national service like this, which, like the said, has a lot of diversity in its users and so on.

Phil Ewels: Makes a lot of sense.

Steven Manos: Mm-hmm.

Next steps

Phil Ewels: So, for Australian and or so non Australian bioinformaticians users. What should they go and click now? What, what do you, what do you want them to do next? What’s your take home for everyone listening?

Ziad Al Bkhetan: Well, I would say for Australian researchers, you know, please check out our website. You’ll see a list of all the services we have, with some contact details in case some of these services are helpful for you. But also feel free to reach out to us and we can discuss, have a chat, and see how we can help you and facilitate your work, as well.

Steven Manos: Yeah, we’ve also got some great resources online. we have some, various webinars around, Nextflow and Seqera, which are recorded and available on YouTube. They can all be accessed through the BioCommons.org.au website. So I’d encourage you to look at those.

We have our Nextflow community. We’re about to begin an office hours collaboration, as well. So we’re gonna be publishing details of that soon. Please sign up to our newsletter is the very best way to, get notifications, in your email inbox around, what’s coming up in terms of training, changes to services and so on and so on.

Phil Ewels: Fantastic. And we’ll, we’ll put links to all of these things into the, the podcast show notes as well. So links to your website. The, the nf-core Weekly office hours now has a website as well with details of, of when to join that call. So anyone who’s listening can, can go find those things.

Steven Manos: Great.

Wrap up

Phil Ewels: Fantastic. Thank you so much for joining me today, folks. I know we’re talking about time zones. This hasn’t been the easiest podcast to schedule, but it’s a good case in point.

I really appreciate the time and, and thanks very much and, really impressive work you’re doing in Australia, and I hope it can provide a blueprint for many other initiatives around the world, to see your success and, and build on that. Long may continue.

Steven Manos: Thank you.

Ziad Al Bkhetan: Thanks so much.

Phil Ewels: Alright, thanks all and thanks everyone for listening. We’ll see you the next episode.