HDL cholesterol

HDL is the 'good' cholesterol that helps clear fat from the blood. This polygenic score estimates your inherited tendency toward higher or lower HDL from many common variants.

What this HDL score reads

This is PGS000064 in the PGS Catalog, built by Kuchenbaecker and colleagues and published in Nature Communications in 2019, in a study asking how well lipid genetics carries from one population to another. It sums 120 common variants that reached genome-wide significance for HDL cholesterol across a meta-analysis of 331,368 people, about 82% of them of European ancestry.

Each of those variants nudges HDL by a fraction of a milligram per decilitre. The score is their weighted total, so no single variant decides your result, and no single variant in it is worth reading on its own.

How well it predicts HDL

The authors measured this directly, which many scores never report. In European-ancestry cohorts the score correlated with measured HDL at r = 0.27 to 0.29, meaning it accounts for roughly 8% of the person-to-person difference in HDL. In an East Asian cohort the correlation fell to 0.18, and in a sub-Saharan African cohort to 0.12, where it explains only 1% to 3%.

Read that the other way around: even in the population it works best in, more than 90% of the variation in people's HDL is not captured here. Body weight, alcohol, physical activity, smoking, thyroid function, diabetes control and several medications all move HDL, and none of that is in a DNA file.

How to read a high or low score

Your result is a percentile, a ranking against a reference population, not a measurement. A score in the 90th percentile means your inherited tendency leans toward higher HDL than most people in that reference group, not that your HDL is currently high.

HDL is also one of the cheapest things in medicine to actually measure. A standard lipid panel gives you the real number rather than an estimate that explains under a tenth of the variation, so treat this score as background on why your measured value sits where it does, not as a substitute for it.

What this score does not capture

Rare large-effect variants are missing entirely. Conditions like Tangier disease (ABCA1) and LCAT or APOA1 deficiency can push HDL far outside the normal range on the strength of a single variant, and none of them are read by a consumer genotyping chip or included in this score.

It is also worth being clear about what HDL itself means. HDL tracks with lower cardiovascular risk in observational data, but drugs that raise it, including niacin and the CETP inhibitors torcetrapib, dalcetrapib and evacetrapib, did not reduce cardiovascular events in trials. A high HDL score is not protection you can rely on, and a low one is not a diagnosis.

Ancestry and accuracy

This score was built almost entirely from European-ancestry data, and its own validation shows what that costs: the correlation with measured HDL more than halves in the sub-Saharan African cohort. That pattern is general. Across traits, Martin and colleagues found polygenic score accuracy roughly 1.7 times lower in South Asian, 2.5 times lower in East Asian and 4.9 times lower in African-ancestry individuals than in Europeans, because prediction decays with genetic distance from the population the score was trained in.

Variants in this score you can read on their own

This score includes a variant that this site also explains individually, where you can see what each genotype means on its own: APOE type (Alzheimer's & cholesterol) (rs7412).

How much of this score your file covers

All 120 variants in the published score are present in a standard consumer DNA file, so this reconstructs the published score in full.

Common questions

Does a high HDL polygenic score mean my HDL is actually high? Not reliably. The score explains about 8% of HDL variation in European-ancestry people and less in others, so it is a lean, not a measurement. Plenty of people with a high score have ordinary HDL, and a blood lipid panel settles the question directly.

Is a high HDL score good news for my heart? It is weaker news than it sounds. Higher HDL is associated with lower cardiovascular risk in population data, but trials of drugs that raise HDL did not lower event rates, which is why HDL is no longer treated as a target to push upward. LDL cholesterol, blood pressure, smoking and blood sugar carry far more weight in cardiovascular risk.

Why does this score work less well for some ancestries? The variants and weights come from a study population that was about 82% European. Which variants tag a real effect, and how strongly, depends on local patterns of inheritance that differ between populations, so the same score carries less information further from where it was built. The score's own paper measured that drop across African, Asian and European cohorts.

Related

Adult lactose digestion · Alcohol response (flushing & dependence risk) · Beta-carotene to vitamin A conversion (BCO1) · LDL cholesterol · Omega-3/6 conversion efficiency · Secretor status (norovirus resistance & B12) · Triglycerides