Local DNA Explorer · Phase 2 preview
Your raw DNA file, opened on your own machine
Select a raw genotype file from your disk. DNADojo detects the format, reports the reference build, counts what is in it and lets you search — with no upload, no account, and no third-party script on this page.
Select a raw genotype file
IdleStarting…
Search the parsed file
How it works
Four steps, none of which involve a server
Read the header
The first 64 KB are decoded to identify the delimiter, the column layout and any build statement.
Pick an adapter
Each format adapter scores its confidence. The highest score wins, and the score is shown to you.
Stream the file
4 MB at a time through
File.slice()andTextDecoder, so memory stays flat and the page stays responsive.Index in typed arrays
Chromosome, position and genotype are stored columnar, which keeps a 700k-SNP file in tens of megabytes rather than hundreds.
What this preview does and does not do
| Capability | Status | Note |
|---|---|---|
| Format detection | Working | Single-genotype, two-allele and generic delimited layouts |
| Build detection | Working | Reported, never converted — see below |
| Streaming parse with cancel | Working | Cancellable at any point; memory is released |
| rsID and region search | Working | In-memory index, no persistence |
| Saving a project on this device | Not enabled | IndexedDB is wired but opt-in and off by default |
| Matching against other people | Never | DNADojo does not build a cross-user DNA database |
| Health interpretation | Never | Out of scope by design |
Reference builds are reported, not converted
GRCh37 and GRCh38 number the same physical positions differently. Converting between them reliably needs a proper lift-over with a clear record of what was converted and what failed. Until that exists here, the Explorer shows you the build — including when the file does not state one — and refuses to guess.