Equip DNA Info Lab · Revisió editorial · · 6 min de lectura
23andMe Raw Data Analysis: What You Can (and Cannot) Learn
What can a 23andMe raw data analysis tell you? Research associations from the GWAS Catalog, pharmacogenomics (drug response), trait associations and some carrier-status hints — all from common variants. It cannot give a diagnosis, detect rare mutations not on the chip, or predict your future. Here is the honest scope.
Once you have your raw file, the useful question is not "what can I run" but "what can the result actually mean". Consumer microarray data is powerful for some questions and useless for others. Here is the honest boundary.
What you CAN learn
- Research associations — which of the variants studied for common conditions (type 2 diabetes, cardiovascular, Alzheimer's, depression and more) are in your file, each with the trait its study measured and the effect it reported, from the GWAS Catalog. We do not add them into a score: see our methodology for why.
- Pharmacogenomics — how your genotype may affect response to common drugs (warfarin, statins, clopidogrel, codeine, SSRIs, metformin), from PharmGKB.
- Trait associations — well-studied, low-stakes traits like lactose tolerance, caffeine metabolism, or the ACTN3 "sprint" variant.
- Carrier hints — some ClinVar-flagged variants that are on the chip. Not a full carrier screen, but a starting signal.
What you CANNOT learn
- A diagnosis. An association measured across a study population is not a verdict about you. Carrying the studied allele is not disease; not carrying it is not safety.
- Rare private mutations. If a variant was not on the chip (~99.98% of your genome isn't), it simply is not in the file. A "clear" result never rules out a rare mutation.
- Your full genome. Microarrays read ~650,000 chosen positions. For everything else you need whole-genome sequencing.
- The future. Genetics is one input among many — lifestyle, environment and chance dominate most outcomes.
The ancestry caveat
Roughly 90% of the participants in the studies behind these associations are of European ancestry. If your ancestry is different, the effects they report may not transfer with the same magnitude — an honest analysis flags this rather than hiding it. Treat non-European results with extra caution.
How accurate is it?
The genotyping itself is highly accurate — over 99% concordance with sequencing at the positions the chip measures. The uncertainty is in interpretation: effect sizes come from population studies with confidence intervals, and how much of an area your file covers varies (a category the chip matched at 40 positions was checked against more research than one matched at 4 — which is a statement about your file, not about how much to trust either result). Good tools show you the coverage and the source for every match, and no aggregate figure they cannot defend.
How to run the analysis
You can do it by hand — look up individual rsids in SNPedia and the GWAS Catalog — or upload the whole file to a tool that cross-references it automatically and explains each result, with its limits, in plain language.
Frequently asked questions
Can 23andMe raw data diagnose a disease?▼
No. It reports associations measured across research populations, not a clinical diagnosis and not a prediction about you. Anything concerning should be confirmed with a doctor and, where appropriate, clinical-grade testing.
Will an analysis find every genetic condition I might carry?▼
No. It only covers variants on the chip (~0.02% of the genome). A reassuring result never rules out a rare mutation that was not tested.
Does non-European ancestry change what the analysis can say?▼
For the GWAS research areas, yes: those studies over-represent European-ancestry participants, so an effect measured there may be different, or absent, in your population. It does not apply the same way to the clinically-reviewed ClinVar findings or the drug annotations, which are not filtered by ancestry. Interpret the research areas with caution.
Do I need whole-genome sequencing instead?▼
Only if you need rare-variant coverage. For common-variant risk, pharmacogenomics and traits, microarray raw data is more than enough.
Run your analysis
Upload your 23andMe, AncestryDNA or MyHeritage file and see your clinically-reviewed ClinVar findings and strong-evidence drug interactions free. The full report — ten health areas of research, with the trait each study measured and its source — is 10 €.
Related articles
Aquest article té finalitats exclusivament educatives i no constitueix consell mèdic, diagnòstic ni tractament. Consulta sempre un professional sanitari per a decisions sobre la teva salut.