Pipelines that reproduce
Workflows that give the same answer after the person who wrote them has left. Tool versions, parameters and intermediates are recorded, so the numbers can be regenerated months later.
Consulting
We start where your data stopped
From raw sequencing reads to the figure in your paper, we find the point where the work actually stalls and fix that part. We look at the problem, not at what we could sell you.
What we build
Every lab gets stuck somewhere different. These four come up most.
Biomarker discovery
Statistics and machine learning to narrow candidates, written up so you can say why each one survived. Validation design is part of the conversation.
Integrating unlike data
Omics, clinical and imaging data joined into one cohort. Standardisation rules and missing-data handling are documented, so whoever joins later makes the same calls.
Structure-based design
A design loop from target structure to candidate sequences. We do not take predicted scores at face value; we agree with you on what has to be checked at the bench.
How work usually starts
Not with a large contract. Small first, wider once it proves useful.
-
01
Find where it stalls
An hour or two of conversation. The problem is usually not the method but the handoffs around it.
-
02
Build one narrow piece
The most painful part first, before designing the whole. Typically two to four weeks.
-
03
Run it on your real data
Not an example set. This is where most of the assumptions get corrected.
-
04
Integrate, or stop
If it earns its place, it gets wired into your environment. If not, we stop there — that is also a result.
What we offer
Pipeline development
Design and build for WGS, WES, RNA-seq, scRNA-seq and related data types.
Review of existing analysis
An outside read of the methods and interpretation behind work you have already run.
Standardisation and integration
Scattered data brought into an analysable shape, with the rules written down.
Applying AI models
Choosing and validating predictive or in-silico models against your actual question.
Interpretation and reporting
Turning numbers into sentences and figures that can go into a paper or a report.
Platform onboarding
Setting up AURORA, BioWrit or PepDesigner to fit how your group works.
Training and handover
Making sure your team can run what was built without us.
Who we work with
University and hospital labs, research groups inside biotech companies, and small teams that cannot justify a full-time analyst. The most common starting point is a group with data piling up and no clear next step.
Common questions
Can we talk before we have data?
Yes, and it is usually better. Most analysis problems are decided at the experimental design stage.
How long does it take?
A first narrow deliverable is typically two to four weeks. Full pipeline work depends on scope, which we set together after the first conversation.
Does our data have to leave the institution?
No. Work can be set up to run inside your environment, and we default to that because clinical data usually cannot be exported.
Who owns what gets built?
You do. Code and documentation are handed over so your team can run it.
What kind of project fits best?
Specific ones. "This analysis gives a different answer depending on who runs it" resolves far faster than "we want to adopt AI".
Have a problem in mind?
Send us a paragraph on what is stuck. If it is not a fit, we will say so.