https://nf-co.re logo
Join Slack
Powered by
Hi todos, i am Allyson a PhD student who want to c...
# say-hello
a
Hi todos, i am Allyson a PhD student who want to create a new nf-core pipeline and I am here to speak to the community before doing it as suggested by nf-core !
👋 5
👍 1
👋🏻 1
m
Welcome to nf-core, which kind of pipeline would you like to add to nf-core?
a
a pipeline of detection and quantification of ribosomal RNA variants for long read
m
Sounds interesting, I don't think we have long reads in #C05785S3TMX, do not hesitate to join that channel, or even #C0B5DGK8AQG that could be of interest for you
a
ok thanks, so I understand that my pipeline need to be accepted before I start it ?
m
I do recommend to start the proposal as soon as possible, that way we can avoid any duplication of work, and maybe even help fostering collaborations
a
ok I do it right now
m
Thank you so much
a
thank you for the welcoming
I have a question. Is my request public in the new pipeline proposal process of nf-core ? Because I don't want my shematic diagram being public while my code is not finished
m
All requests are public indeed
a
Ok so maybe I can create a custom pipeline and then when my thesis is published I would request a new pipeline proposal ? It should work ?
p
Hi @allyson.moureaux - you can do that, and some do. But to do so kind of misses the point of nf-core. We are not a registry of pipelines. We are a community of people who collaborate on building pipelines.
The flow is designed to foster collaboration.
Your path has several risks: someone else may propose the same pipeline first, in which case your proposal and pipeline will be rejected when you come to propose it.
Also coming in with an already-finished pipeline risks having significant and fundamental feedback and requests at a late stage, which can be much more work
Ideally you propose the idea, and build out the plan with the community. Then start by contributing the individual modules (each of which is reviewed independently as you go). And finally assemble them into a pipeline.
Hopefully you gain interested parties along the way and others can help with testing and development of the pipeline
See for example #C0BVDK00QKY - a new pipeline that was accepted this week where the development is happening in the open with the community
a
Ok i see that I missed the main purpose of nf-core. I understand now. I need to speak to my PI before opening my code to everyone
p
No worries! Shout if you have any questions 🙂
If you need help convincing: it is this style of development that is the main driver of nf-core's success. It means that pipelines typically have a mini-community of people involved in their development and use, which leads to much better testing and quality and long-term maintenance.
Much of this is discussed in our most recent paper: link.springer.com/article/10.1186/s13059-025-03673-9
And you can see it in action in the older pipelines, where development changes hands over the years as different people come in and out: ewels.github.io/contributor-graphs/rnaseq.html
a
Ok I see. I will discuss with my two PI and see what comes out. I reach back soon if we go for it !
thanks for all the informations
p
No problem! Thanks for reaching out 🙂 Much better to have these conversations early instead of after months of work! 😅
a
I see what they will ask me : If I start a pipeline in nf-core do I choose what I want to do in it and the changes needed through my work or the idea is not mind anymore and I can't change my process how I want ?
p
You will still own the work that you do - you're free to drive the direction and development, publish what you build and so on.
In practice you'll probably find that you're the only developer, or at least the main developer - with others dropping into add minor features or fix bugs
So it's extremely unlikely that anyone is going to come in and suddenly change what you're trying to build, if that's the concern
We do have a rule that you need to work with the community - so you can't just ignore folks basically 😅 If there is disagreement (very rare) we generally take community consensus, and the core-team makes a decision on how to proceed. This has happened maybe once or twice ever in the 8 years we've been running (with 150+ pipelines now).
You can see that many people have successfully published papers and PhDs with nf-core pipelines: nf-co.re/publications
a
Yes that is the concern ahah I will maybe create a process and then erase it because I don't want that anymore in my pipe
ok I see so if some publish paper with it I think it would be ok
p
yup absolutely, better than ok - definitely encouraged 🙂
😉 1
We do say that pipelines are "community owned". So basically you agree that nf-core as a community ultimately owns and controls the pipeline. This is so that if you get a different job and disappear, others in the community are free to continue working on the pipeline and potentially change direction etc.
But yeah, in practice we have basically never had any problems like you're saying, where other people make changes that are unwanted by you.
👍 1
a
I see it avoid 3 years of coding for no future after changing subject in a post doc !
p
Exactly 💯
s
Yes, there are so many pieces of bioinformatics software that I call "PhD abandonware". They have been written as part of a PhD, and then never touched again. No long-term maintenance, no fixes for anything after they were published. And that makes them hard to use and hard to maintain. The same applies for pipelines, if you write something just for your use then it is likely to only ever be used for yourself (and maybe people after you in the same lab). Whereas if you have a community of people working on it, then it can become bigger than that. Some of our core pipelines were first laid down 10 years ago, and have gone through a number of people maintaining them as well as using them.
💯 1
2 Views
Open in Slack
PreviousNext
Powered by