Your genome, decoded in the browser.
Drop in a 23andMe or AncestryDNA export. DNA Genie scans 600,000+ markers locally, keeps only the 44 that matter, and explains what they mean — with the studies to back it up.
- 40
- traits explained
- 44
- SNPs read from your file
- 0 B
- of raw DNA uploaded
- Aadenine
- Tthymine
- Gguanine
- Ccytosine
Interactive
Walk the helix.
Every marker on this strand is a real SNP the app reads from your file. Orbit around it, filter by category, and fly into any trait to see what it can tell you.
How it works
Three steps. Nothing leaves your tab.
The whole pipeline — detection, parsing, interpretation — runs client-side. The server only ever sees a login.
- 01
Export your raw data
Download the raw genotype file from 23andMe or AncestryDNA. It's a plain text file with ~600,000 rows — one per marker.
# This data file generated by 23andMe# rsid chromosome position genotypers548049170 1 69869 TTrs13328684 1 74792 --rs9283150 1 565508 AA… 600,000 more rows - 02
Parsed on your device
A streaming parser reads the file in 10,000-line chunks in your browser tab. Only the 44 trait-relevant rsIDs are kept; everything else is discarded in memory.
parser · chunk 1/61rs4680chr22GG✓ keptrs762551chr15AA✓ keptrs1815739chr11CT✓ keptrs9939609chr16AT· skiprs12913832chr15GG✓ kept - 03
Read your traits
Each genotype is matched to a curated interpretation: what it's associated with, practical tips, and the PubMed studies behind it.
CYP1A2Caffeine metabolismFast metabolizerPER2ChronotypeNight owlACTN3Muscle fiber typePower / sprint
The panel
40 traits, 8 categories, every one cited.
No polygenic scores or black-box risk numbers — just well-studied single variants with plain-language interpretations.
Privacy by architecture
0.007% of your file. That's the whole footprint.
Most consumer genetics tools ask you to upload the entire file. DNA Genie was designed so it can't.
Your raw file
Stays on disk. Read once by the browser with the File API.
→In-browser parser
Streams line by line, keeps only the markers in the trait panel, discards the rest in memory.
→Your account
Only if you sign in. Firestore rules scope every document to your own user ID.
Nothing to leak
There is no upload endpoint. The raw file never becomes a network request, so there's no server log, bucket, or backup holding it.
Demo without an account
Try the sample genome, or parse your own file in demo mode — results live in session storage and vanish when you close the tab.
Open source
The parser, the trait database, and the Firestore rules are all public. Read exactly what runs.
talal-nabulsi/dna-genie ↗FAQ
Straight answers.
Genetics is easy to overstate. Here's where the lines are.
Raw data exports from 23andMe (.txt) and AncestryDNA (.txt or .csv). The parser auto-detects the format from the header, normalises strand orientation where needed, and reads genotypes as unordered allele pairs.
Ready to meet your genome?
Start with the sample genome — no account, no upload. Bring your own file whenever you're ready.