What to Do With Your 23andMe Raw Data File
You requested your raw data from 23andMe, waited for the download, and now you have a large file you cannot really read. This guide explains what is in it, what it can and cannot tell you, and how to explore it without handing it to anyone else.
What the raw data file is
A raw data file is a plain text list of the genotypes that a genotyping chip read from your saliva sample. It is usually a compressed archive that contains one text file. Open it and you will see a few comment lines at the top, followed by rows with four tab-separated columns:
- rsID: an identifier for a known variant, such as rs followed by a number
- Chromosome: which chromosome the variant sits on
- Position: where on that chromosome it is located
- Genotype: the two letters you carry at that spot, one from each parent
A typical chip reads roughly 600,000 to 700,000 of these markers. That sounds like a lot, but the human genome has about three billion positions. The file is a sparse sample of well-studied places, not your whole genome.
What the file is not
It helps to set expectations before you start.
It is not a full sequence
Chips look only at variants that were chosen in advance. Most rare variants are not on the chip, so they are simply absent from your file. The absence of a variant in your file does not mean you do not carry it.
It is not perfectly accurate
Genotyping chips can misreport some variants, especially rare ones. A surprising result at a single marker is a reason for caution, not for conclusions.
It is not a diagnosis
A published association between a variant and a trait is a statistical finding in groups of people. It is not a statement about what will happen to you. Most traits are shaped by many variants, by environment and by habits.
Why privacy matters from the first step
Your genome is the one piece of personal data you cannot change. It also says something about your relatives. Before you upload the file anywhere, ask three questions:
- Where does the file go once I load it?
- Do I need an account, and what is stored under it?
- Are my questions and the answers kept by someone else?
Many online tools want the whole file on their servers. You do not have to accept that trade to learn something from your data.
A practical way to start
Here is a simple, careful routine.
Step 1: Keep a clean copy
Store the original download somewhere you control, such as an external drive or a folder on your own computer. Avoid emailing it or placing it in shared folders.
Step 2: Look at the file yourself
Open the text file in a plain text editor if you are curious. Seeing the rows makes the idea of a genotype concrete. Do not edit it.
Step 3: Pick a question, not a feature list
Good questions are narrow and educational. For example: how is caffeine metabolism linked to common variants? Or: what do researchers say about lactose tolerance in adults? Starting with a question keeps you away from scrolling through hundreds of unrelated results.
Step 4: Check the sources
Any explanation should point to research you can read yourself. Useful public resources include PubMed for published papers, NCBI ClinVar for curated variant reports, the NHGRI-EBI GWAS Catalog for association studies, and 1000 Genomes for how common a variant is across populations.
Step 5: Treat results as education
If something in your results worries you, talk to a qualified health professional or a genetic counselor. A raw data file is not a clinical test.
Exploring your file with DNA2C
DNA2C is built for exactly this situation. It reads raw data files from 23andMe and AncestryDNA, and you ask questions in plain language.
The important part is where the work happens. The AI model runs on your device. Your genome, your questions and the answers are never sent to us or to any AI company. There is no account to create.
To cite sources, DNA2C looks up public research databases using only variant IDs and gene names, never your genotypes. That means the databases learn which public variant was asked about, not what you carry.
There is also a built-in sample genome, so you can try the app and see how answers look before you load your own file. Example topics include caffeine, lactose tolerance, endurance versus sprint, nutrition and exercise.
On Mac
On the Mac, data is stored encrypted on your Mac. You get three questions free, and DNA2C Pro is a one-time in-app purchase with no subscription.
On Android
On Android, DNA2C is a paid download, pay once, with no subscription and no in-app purchases. The AI model runs on your phone, and your DNA file, genotypes and the AI's answers stay on your device. There are no ads and no tracking. It needs a recent 64-bit Android phone, with 4 GB of RAM or more recommended, and it downloads its AI model once, about 800 MB.
A short checklist
- Keep the original file safe and offline where you can
- Prefer tools that work on your device
- Ask narrow questions
- Look for sources you can open and read
- Remember that an association is not a diagnosis
Education only
DNA2C and this article are for education only. They are not a medical device and do not provide medical advice or diagnosis. Talk to a health professional about any health concern.
Your 23andMe file holds a lot of curiosity value, and you can enjoy it without giving it away. DNA2C on Mac and on Android is designed so that you can ask questions about your raw data, read the sources, and keep the file on your own device.
