Population differences beyond polygenic scores

The EDAR variant associated with thicker scalp hair and shovel-shaped incisors occurs on 93.7% of chromosomes in the Beijing Han Chinese sample of the 1000 Genomes Project. In its European sample, the frequency is 1.1%. A different variant, in ALDH2, helps explain why a small amount of alcohol can cause intense facial flushing in some people. These are familiar human differences with identifiable genetic contributors. [13, 15, 18, 19]

Much of my research concerns polygenic traits. Height, educational attainment and many other outcomes reflect the combined influence of large numbers of genetic variants, usually with individually small effects. Polygenic scores summarize some of that variation. Their appeal is that they let us study an outcome whose genetic basis is spread across the genome. Their interpretation depends on how the score was built, the people in whom it was tested, and the environmental conditions connecting genotype to phenotype. [1, 2, 3]

For this post I wanted to look more closely at a few particular changes. Some alter one DNA letter, a single nucleotide polymorphism or SNP. Others remove an entire gene or give a person extra copies. We can ask what an enzyme does, how a receptor develops hair and teeth, or how a red blood cell becomes harder for a parasite to infect. The biological questions are often more specific than those addressed by a genome-wide score.

I chose these examples because they combine substantial population variation with functions worth understanding: starch digestion, androgen metabolism, defence against parasites, and aspects of appearance. They also differ in how securely we can connect the DNA change to a physical outcome. UGT2B17, the testosterone-metabolizing gene, is especially instructive. Its deletion has a large effect on urinary hormone excretion, while its relationship with body fat and muscle is much less straightforward. [7, 8]

The plots make the geographic differences explicit. Figure 1 starts with four single-letter variants in a consistent reference panel. The following sections examine the variation within Asia, the ancient history of EDAR, and the deletions and duplications that a conventional SNP analysis can overlook. These are selected case studies, so their striking contrasts should not be taken as a measure of differentiation across the genome as a whole.

Figure 1. Four variants with contrasting population distributions

EDAR and the genetics of East Asian appearance

EDAR is often described as an East Asian appearance gene because one of its variants affects several features common in East Asians. V370A, or rs3827760, increases signalling through a receptor involved in the development of hair, teeth and glands. Human studies associate it with thicker, straighter scalp hair and shovel-shaped incisors. Facial studies also associate it with reduced chin protrusion. Experiments in mice support direct effects on hair thickness and gland development. [13, 14, 21]

It is one contributor to East Asian appearance, not a single explanation for it. Facial form and hair characteristics themselves involve many genes. EDAR is interesting precisely because one functional variant has several effects that can be studied individually.

The illustration below highlights a connection across the Pacific: EDAR V370A is frequent in many East Asian and Indigenous American populations and contributes to shared hair and dental traits. [13]

An EDAR link across the Pacific

Upgrade to explore the population-by-population plots, trace EDAR through ancient genomes, and find out whether deleting UGT2B17 changes testosterone, body fat or muscle.