Asthma

Asthma is a common chronic airway condition with a real but diffuse genetic component. This score sums 22 common variants, and the study it comes from found that scores like it did not improve asthma prediction in children.

What this asthma score reads

This is PGS000799, a 22-variant score used by Dijk and colleagues in 2019, drawn from the large TAGC asthma meta-analysis of 169,861 people, about 87% of European ancestry.

The paper it comes from is unusually blunt, and its title states the conclusion: genetic risk scores do not improve asthma prediction in childhood.

How well it predicts asthma

In two birth cohorts, each standard deviation higher score carried odds ratios of 1.23 and 1.52 for asthma by age eight. Prediction models reached areas under the curve of 0.66 and 0.69, but only when the genetic score was combined with family history, environmental exposures and perinatal factors, and the authors found the genetic component added little to what those already provided.

That is the finding to carry away from this page. The score detects a real signal at group level and does not usefully predict asthma in an individual child.

An important gap in what can be read

Six of the score's 22 variants cannot be read from a consumer DNA file, and they include the three largest-weighted variants in it, which sit in the IL13 and IL33 pathways central to allergic airway inflammation.

A result computed from a consumer file is therefore not just an approximation of this score, it is missing its strongest ingredients. Treat any percentile from it as weaker than the already weak published performance implies.

What this score does not capture

Asthma is shaped heavily by exposures: respiratory infections in early life, tobacco smoke, air pollution, allergens including dust mite and pets, occupational exposures, obesity and, in some people, aspirin sensitivity. None of these are genetic and all matter more for an individual than this score.

Ancestry and accuracy

The source meta-analysis was about 87% European ancestry and the validation cohorts were European-ancestry children. Asthma prevalence and triggers vary substantially across populations and environments, and this score has not been shown to carry information outside the setting it was tested in.

How much of this score your file covers

16 of the published score's 22 variants are present in a standard consumer DNA file (73%). The remaining 6 cannot be read from one, so a result computed here approximates the published score rather than reproducing it.

Common questions

Can a DNA test tell me if my child will develop asthma? Not usefully. The paper behind this score tested exactly that question in birth cohorts and concluded that genetic risk scores did not improve childhood asthma prediction over family history and environmental factors.

Why does my asthma score matter so little when asthma runs in families? Family clustering reflects shared environment as well as shared genetics, and the genetic part is spread across many variants plus rare ones that this 22-variant score does not read. Familial risk is captured better by asking about your relatives than by this percentile.

Related

Age-related macular degeneration · Age-related macular degeneration risk · Coronary artery disease risk (9p21) · Diastolic blood pressure · Prothrombin clotting variant · Systolic blood pressure · Uric acid / gout tendency