Skip to main content
📚 Science

Methodology & Scientific Basis

How we read your genetic data, and the limits of what it can support.

1. Research associations (GWAS)

Common diseases rarely depend on a single gene: they arise from many genetic variants — hundreds or even thousands per disease — together with environmental factors. We use the GWAS Catalog (Genome-Wide Association Studies), run by NHGRI and EMBL-EBI, to find which of those variants are present in your file, and for each one we show the trait the study measured and the effect it reported — an odds ratio or a beta. The catalogue lists associations with p < 1×10⁻⁵; we keep only those at genome-wide significance (p < 5×10⁻⁸), the standard threshold for common-variant studies. Because GWAS participants are mostly of European ancestry, we use only analyses recorded as European-ancestry, and the report says so; when one study published several analyses of the same association, we set them all aside rather than choose one. We do not add those effects into an overall score. Our own audit found that more than half of a category's combined effect came from studies of something other than the condition it names, and that betas were being summed across units that are not comparable, so the aggregate did not measure what its label said.

Sources: NHGRI — polygenic risk scoresGWAS Catalog — aboutGWAS Catalog — eligibility criteriaFadista et al. 2016Fatumo et al. 2022

2. Pharmacogenomics (PharmGKB)

We show drug–gene annotations from PharmGKB, now part of ClinPGx. Its curators grade each clinical annotation by level of evidence, from 1A (variant-specific prescribing guidance in a clinical guideline or an FDA-approved drug label) to 4 (the evidence does not support an association). We include levels 1A to 2B for every drug, and lower levels only for a short list of commonly prescribed medicines. An annotation describes how a genotype has been associated with a drug's efficacy, toxicity, dosage or metabolism; it is not a prescription, and no medication should be changed without your doctor.

Sources: ClinPGx — levels of evidence

3. Clinical Variants (ClinVar)

ClinVar, maintained by NCBI, is a freely available public archive: clinical testing laboratories, research laboratories, expert panels and other groups submit their interpretations, and ClinVar aggregates them. This report uses GERMLINE classifications. When the source provides condition-specific assertions (RCVs), each classification is attributed to its condition; when it provides only a variant-level aggregate, the report identifies it as such and lists related names separately. A conflicting classification, if shown, is labelled as conflicting. This is not a diagnosis: confirming one needs genetic testing in a clinical setting, which SNP-chip data is not.

Sources: Landrum et al. 2018

🛡️ Privacy & Security

Your file is uploaded over HTTPS, processed in memory and discarded: what we store are the extracted variants and your report, on our server in Germany (EU). We never sell them or hand them to insurers, employers or advertisers. If you buy the report, it is kept while your account is active; if you do not, the analysis is deleted automatically 30 days after upload, and you can delete it yourself at any time. Two third parties are involved: Cloudflare, our network, carries the upload in transit, and if you request an AI explanation the derived results — rsIDs and your genotypes at those variants — are sent to Anthropic, never the file.

Traits & Conditions Analysed

✔Type 2 Diabetes
✔Insulin Resistance
✔Glucose Metabolism
✔Lipid Profile (Cholesterol/Triglycerides)
✔BMI & Obesity Predisposition
✔Inflammation Markers
✔Cardiovascular Health
✔Alzheimer's Risk
✔Major Depression
✔Common Cancer Predisposition