Six of these positions shift the odds slightly and mean nothing on their own. The seventh is different in kind, and it is the reason this page spends more space on what a result cannot tell you than on what it can.
Parkinson's disease is a movement disorder caused by the loss of dopamine-producing neurons in the substantia nigra, a small structure deep in the midbrain. Tremor, slowness and rigidity appear once a substantial fraction of those cells are already gone, which is why the diagnosis arrives years after the process starts.
It was long described as the environmental disease — the one caused by something you were exposed to rather than something you inherited. That turned out to be half right.
A 2011 meta-analysis of five genome-wide studies looked at 5,333 cases against 12,019 controls, then repeated the exercise in an independent 7,053 cases and 9,007 controls. Eleven positions passed genome-wide significance. Two of them — MAPT and SNCA — were already known; the rest were new.
A different study the same year recruited its participants online, reaching 3,426 cases and 29,624 controls. It found rs6812193 near SCARB2 and replicated twenty earlier findings, and it put a number on the whole question: heritability at least 0.27.
So the environmental account was incomplete. Inheritance matters here, and rather more than the textbooks said.
Six of the seven variants on this page are ordinary common-variant findings. They shift risk by a few percent, they are carried by large fractions of everyone, and no combination of them tells an individual anything useful.
rs34637584 is a different object. It is LRRK2 G2019S — a single amino-acid change, not a marker near a gene but the change itself, and the most common known genetic cause of Parkinson's disease.
It is rare in most populations and much less rare in others: it accounts for a substantial share of cases among Ashkenazi Jewish and North African Arab populations, and very few elsewhere.
Carrying G2019S does not mean developing Parkinson's disease. Published estimates of lifetime risk in carriers vary widely between studies and populations, and the honest summary is that a majority of carriers do not develop the disease — but the risk is real and it is far above the population baseline.
That combination — clearly elevated, clearly not certain, and not estimable precisely for an individual — is the hardest kind of genetic result to receive.
There is no treatment that prevents Parkinson's disease and no test that says when or whether it will start. A person who learns they carry G2019S gains no action they can take and no date they can plan around.
This is why predictive testing for adult-onset conditions with no preventive treatment is done through genetic counselling rather than through a saliva kit. The counselling is not a formality: what is being decided is whether to know something that cannot be acted on, and that decision is different for everyone.
If you already have a genotype file that shows this variant, or Parkinson's disease runs in your family, the useful next step is a neurologist or a clinical geneticist. Not this page.
Sources. International Parkinson Disease Genomics Consortium (Lancet 2011) meta-analysed five GWAS datasets from the USA and Europe: discovery of 5,333 cases and 12,019 controls with genotyped and imputed data at 7,689,524 SNPs, replication in 7,053 cases and 9,007 controls, identifying 11 loci past genome-wide significance. Do et al. (PLoS Genet 2011) conducted a web-based case-control GWAS of 3,426 cases and 29,624 controls, reporting rs6812193 near SCARB2 (p = 7.6 × 10⁻¹⁰, OR 0.84) and rs11868035 near SREBF1/RAI1 (p = 5.6 × 10⁻⁸, OR 0.85), both replicated in an independent cohort, alongside replication of 20 previously reported associations including LRRK2, GBA, SNCA, MAPT, GAK and the HLA region.
Heritability and prediction. Do et al. estimated the heritability of Parkinson's disease at at least 0.27 using genome-wide sharing between distantly related individuals, and reported sparse-regression predictive models accounting for 6–7% of the variance in liability — a figure the authors themselves frame as indicating a substantial but far from total genetic contribution.
LRRK2 G2019S (rs34637584). A missense substitution in the kinase domain of leucine-rich repeat kinase 2, and the most common known monogenic cause of Parkinson's disease. Frequency is strongly population-dependent, being appreciably higher in Ashkenazi Jewish and North African Arab populations than in most others. Penetrance is incomplete and age-dependent, and published estimates differ substantially between ascertainment designs; family-based series generally give higher figures than population-based ones. It is not equivalent in kind to the common-variant associations elsewhere on this page and should not be interpreted alongside them as though it were.
Clinical use. No disease-modifying therapy exists, and no genotype currently alters management. Predictive testing in an asymptomatic person is a counselled decision, not a laboratory one. Diagnosis remains clinical, supported where necessary by dopamine transporter imaging.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Parkinson's Disease comes down to these specific, well-studied positions — not a diagnosis.
That the risk is well above the population baseline and still not a certainty — most carriers do not develop Parkinson’s disease. Published penetrance estimates vary widely with study design, and none of them produces a number for one person. There is no preventive treatment, so nothing changes today. Take it to a clinical geneticist or a neurologist rather than to a search engine; that is what the counselling around this variant exists for.
That was the older view and it was incomplete. Genome-wide sharing between distantly related people puts heritability at at least 0.27, and eleven loci passed genome-wide significance in a meta-analysis of more than twelve thousand cases. Environment matters; so does inheritance.
No. The best predictive models built on this data accounted for 6–7% of the variance in liability — informative about populations, useless about individuals. The one variant here that is different in kind, LRRK2 G2019S, raises risk substantially and still does not predict.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.