Call for data owners: bring us a real biological question
Do you have a dataset sitting on a drive and a question about it you have never had time to answer?
We are looking for around 25 researchers — at SciLifeLab, KTH and across the Swedish life-science community — to volunteer that question to someone on the course Data-Driven Life Sciences (SK2538 / FSK3538), which runs at KTH with SciLifeLab for the DDLS community: master’s students, PhD students, postdocs and researchers.
It costs you two 1-hour meetings. You get a working prototype on your data and a written report back.
How it runs
Why we are asking
The 2026 course is built around a single skill: taking someone else’s biological problem, interviewing the person who owns it until the question is precise, and coming back with a working, checked answer — with an AI agent doing the implementation and the student supplying the judgement.
That skill cannot be practised on a tidy teaching dataset, and it cannot be practised on your own data — when you already know the context, you never learn to state it. It needs a real person with a real problem who has not explained it before. So for the final project, every student needs a client. That is what we are asking you to be. We describe the role in full on The Forward-Deployed Scientist.
What we ask, and what you get
How it goes: sign up in about two minutes — an assistant asks a few questions, there is no form · we match you with a participant · a 1-hour meeting where they interview you, and where you hand the data over · they build it over two or three days · a second 1-hour meeting where you say what is wrong · hand-over. Further rounds only if you and the student both want them.
You get a written report and a small analysis app or code that runs on your data after they are gone, and an outsider asking careful questions about your data — which is often worth as much as the analysis. No cost, no obligation, and no claim on your data or results.
You need no programming or AI experience. You bring the biology and the judgement about whether an answer is believable.
What makes a good project
- A real, unanswered question — not a demo, and not something already solved.
- Data that exists now and that you are allowed to share. A de-identified or subsampled extract is completely fine; if nothing can leave your machine, tell us and we will see whether the project can be arranged around that.
- Scope a motivated student can get somewhere with in under three weeks of part-time work. Imaging, omics, screens, clinical tables, sequencing, structures, awkward spreadsheets — all welcome. It does not need to be big data.
- You do not need to know how it should be analysed. Not knowing is the normal case.
Every project gets an answer
We would rather have more projects than people. If nobody is free to take yours, our AI data scientist runs it instead — it interviews you in the same meeting, does the same work, and a member of the teaching team checks everything before it reaches you. Nobody who volunteers goes away empty-handed.
Timeline
| Date | |
|---|---|
| Friday 25 September | Deadline to register a project |
| Late September | We match you and introduce you by email |
| Early October | Meeting 1 |
| October | Prototype, then meeting 2 and iteration |
| By 23 October | Hand-over |
Register
Registering takes about five minutes and gives you a private project page: it is where you upload your sample file, watch the project move through its stages, and raise anything that comes up while it runs.
Not sure whether your problem fits? Write to ddls-course@scilifelab.se. If you have a dataset and a question you care about, it very probably fits.
