See your DNA the way a lab would.
DNA2C Labs is an experimental edition of DNA2C for power users. Walk from a whole genome down to a single letter in 3D, follow how DNA passes through a family, and ask a DNA tutor that runs on your own graphics card.
The prototype runs in your browser on a synthetic sample genome. Nothing is uploaded.
Six ways into one genome
Every screen in the prototype is built on real public data: the DNA sequence and gene map from Ensembl, chromosome banding from UCSC, and a protein structure from AlphaFold. The example follows one well-studied letter in the ACTN3 gene, where a single change turns an amino acid into a stop signal.
Helix Lab
A rotating 3D double helix built from the real DNA around one letter. Switch between the two chromosome copies and watch a codon become a stop.
OPEN →Genome Wheel
All 24 chromosomes drawn to scale with their staining bands, and well-known trait markers like eye colour and bitter taste placed where they sit.
OPEN →Variant Lens
Zoom from chromosome to gene to exon to letters, compare both copies side by side, and preview the local DNA tutor.
OPEN →Protein View
The predicted 3D shape of the protein the gene builds, coloured by domain or by AlphaFold confidence, with the part an early stop removes faded out.
OPEN →Family & inheritance
Simulate a thousand children from two parents, run meiosis again and again, and see why siblings share about half their DNA but never exactly half.
OPEN →Guided lessons
Short chapters that use the genome on screen as the example, with an animated reading frame and quick checks along the way.
OPEN →An AI that reads DNA, running on your own graphics card
We are exploring Carbon-X, an open genomic language model published by HuggingFaceBio on Hugging Face. Unlike a general chatbot, it reads DNA letters and English together, so it can describe what a stretch of sequence is likely to do in plain language.
- Optional download for capable PCs. The rest of Labs works without it.
- Runs locally. Your DNA is never sent to an AI service.
- Every answer is labelled as model output and shown next to facts from public databases.
- Exact work like positions, letters and protein changes is done by ordinary code, not by the model.
Built for people who want to look closer
Be first to try it
DNA2C Labs is an exploration, not a product yet. There is no release date and no price. Join the list and we will email you when early access opens, and maybe ask what you would like to see.
- One email address, nothing else required.
- Stored in the EU. Used only for DNA2C Labs news.
- Leave any time: write to info@yambacode.com.
Frequently asked
Is this medical advice?
No. Labs is for education, curiosity and exploration. It does not diagnose, predict disease risk or recommend treatment. Talk to a doctor or genetic counsellor about anything that matters for your health.
Does my DNA leave my computer?
In Labs, no. Your DNA file stays on your PC, encrypted, and the DNA tutor runs on your own graphics card. The prototype on this site uses a synthetic sample genome and runs entirely in your browser. The waiting list stores only what you type into the form.
How is this different from DNA2C?
DNA2C answers questions about your DNA in plain language, on Mac, Android and the web. Labs is a separate, experimental edition that goes deeper: 3D views, protein structures, family inheritance, lessons and a local DNA model. Both share the same privacy-first core.
What computer will I need?
We are designing Labs for Windows first. The views and lessons are meant to run on most modern PCs. The optional DNA tutor needs a capable graphics card: we expect an NVIDIA card with 12 to 16 GB of graphics memory or more, and will publish exact requirements once we have tested it.
What is Carbon-X?
An open genomic language model published by HuggingFaceBio on Hugging Face under the Apache-2.0 licence. It reads DNA and English. It is a research model, not a clinical tool, and Labs treats its answers as hypotheses to check rather than facts.
When is it coming, and what will it cost?
We don't know yet, and we would rather not guess. Labs is in early exploration. Waiting-list members will hear first.