Height

Adult height is one of the most heritable common traits. This score estimates your genetic leaning toward taller or shorter stature from many common variants.

What this height score reads

This is PGS003837, developed by Zhang and colleagues in a 2023 Nature Genetics paper on multi-ancestry polygenic prediction. It was built specifically in African-ancestry samples: a GWAS of 21,550 people, development in 4,515 and evaluation in 4,527. It contains 49 variants selected by clumping and thresholding.

How well it predicts height, and why the answer is striking

In the authors' African-ancestry validation sample, the score explained 0.3% of the variation in adult height. Essentially none.

That number is worth sitting with, because height is the textbook example of a highly heritable trait: twin studies put heritability around 80%, and the largest height GWAS to date, covering 5.4 million people, found 12,111 variants that together account for about 40% of height variance in European-ancestry populations and roughly 10% to 20% in others.

So height is strongly genetic, well mapped, and this particular 49-variant score still predicts almost nothing. Heritable does not mean predictable from a short list of variants, and a small score built in a smaller GWAS captures a sliver of what is there.

How to read a high or low score

Do not read much into it. At 0.3% of variance explained, the percentile carries no useful information about your height, which you can measure in a doorway.

The classic mid-parental estimate, averaging your parents' heights with an adjustment for sex, outperforms this score by a wide margin, which is a useful reminder of how much information sits in family observation rather than in a genotype file.

What this score does not capture

Childhood nutrition and illness shape adult height substantially, which is why average heights shifted by several centimetres across generations in many countries within a century, far too fast for genetics to explain.

Single-gene and chromosomal causes of short or tall stature, including growth hormone disorders, skeletal dysplasias, Turner syndrome and Marfan syndrome, are clinical diagnoses that a common-variant score does not touch.

Ancestry and accuracy

This score is one of the few here built in African-ancestry data, which was the point of the paper it came from. Even so, it explains very little in the population it was designed for, and less is known about how it behaves elsewhere. It is best read as a demonstration of method rather than a useful personal prediction.

How much of this score your file covers

All 49 variants in the published score are present in a standard consumer DNA file, so this reconstructs the published score in full.

Common questions

If height is 80% heritable, why does this score explain 0.3%? Heritability describes how much of the variation in a population traces to genetic differences in total, including thousands of variants and rare ones. A score explains only what its own variants capture. With 49 variants from a GWAS of about 21,000 people, that is very little, while the largest height study, with 5.4 million people and 12,111 variants, reaches about 40% in European-ancestry populations.

Can a DNA test predict my child's height? Not accurately, and this score in particular predicts almost nothing. Averaging the parents' heights remains the better estimate, and childhood nutrition and health still move the outcome.

Related

Endurance / aerobic capacity (PPARGC1A) · Eye color (blue, green/hazel, brown) · Freckling / sun sensitivity · Hair thickness / shovel-shaped incisors · Male-pattern baldness tendency · Muscle fiber type (power vs endurance) · Red hair / fair skin (MC1R gene)