You need only to look at families to see that height is inherited — and studies of identical twins and families have long confirmed that suspicion. About 80% of variation in height is down to genetics, they suggest. But since the human genome was sequenced nearly two decades ago, researchers have struggled to fully identify the genetic factors responsible.
Studies seeking the genes that govern height have identified hundreds of common gene variants linked to the trait. But the findings also posed a quandry: each variant had a tiny effect on height that together didn’t amount to the genetic contribution predicted by family studies. This phenomenon, which occurs for many other traits and diseases, was dubbed missing heritability, and had even prompted some researchers to speculate that there’s something fundamentally wrong with our understanding of genetics.
Now, a study suggests that most of the missing heritability for height and body mass index (BMI) can, as some researchers had suspected, be found in rarer gene variants that had lain undiscovered until now.
“It is a reassuring paper because it suggests that there isn’t something terribly wrong with genetics,” says Tim Spector, a genetic epidemiologist at King’s College London. “It’s just that sorting it out is more complex than we thought.” The research was posted1 to the bioRxiv preprint server on 25 March.
Scouring the genome
To seek out the genetic factors that underlie diseases and traits, geneticists turn to mega-searches known as genome-wide association studies (GWAS). These scour the genomes of, typically, tens of thousands of people — or, increasingly, more than a million — for single-letter changes, or SNPs, in genes that commonly appear in individuals with a particular disease or that could explain a common trait such as height.
But GWAS have limitations. Because sequencing the entire genomes of thousands of people is expensive, GWAS themselves scan only a strategically selected set of SNPs, perhaps 500,000, in each person’s genome. That’s only a snapshot of the roughly six billion nucleotides — the building blocks of DNA — strung together in our genome. In turn, these 500,000 common variants would have been found from sequencing the genomes of just a few hundred people, says Timothy Frayling, a human geneticist at the University of Exeter, UK.
A team led by Peter Visscher at the Queensland Brain Institute in Brisbane, Australia, decided to investigate whether rarer SNPs than those typically scanned in GWAS might explain the missing heritability for height and BMI. They turned to whole-genome sequencing — performing a complete readout of all 6 billion bases — of 21,620 people. (The authors declined to comment on the preprint, because it is under submission at a journal.)
They relied on the simple, but powerful, principle that all people are related to some extent — albeit distantly — and that DNA can be used to calculate degrees of relatedness. Then, information on the people’s height and BMI could be combined to identify both common and rare SNPs that might be contributing to these traits.
Say, for instance, that a pair of third cousins is closer in height than a pair of second cousins is in a different family: that’s an indication that the third cousins’ height is mostly down to genetics, and the extent of that correlation will tell you how much, Frayling explains. “They used all of the genetic information, which enables you to work out how much of the relatedness was due to rarer things as well as the common things.”
As a result, the researchers captured genetic differences that occur in only 1 in 500, or even 1 in 5,000, people.
And by using information on both common and rare variants, the researchers arrived at roughly the same estimates of heritability as those indicated by twin studies. For height, Visscher and colleagues estimate a heritability of 79%, and for BMI, 40%. This means that if you take a large group of people, 79% of the height differences would be due to genes rather than to environmental factors, such as nutrition.
The researchers also suggest how the previously undiscovered variants might be contributing to physical traits. Tentatively, they found that these rare variants were slightly enriched in protein-coding regions of the genome, and that they had an increased likelihood of being disruptive to these regions, notes Terence Capellini, an evolutionary biologist at Harvard University in Cambridge, Massachusetts. This indicates that the rare variants might partly influence height by affecting protein-coding regions instead of the rest of the genome — the vast majority of which does not include instructions for making proteins, but might influence their expression.
The rarity of the variants also suggests that natural selection could be weeding them out, perhaps because they are harmful in some way.
The complexity of heritability means that understanding the roots of many common diseases — necessary if researchers are to develop effective therapies against them — will take considerably more time and money, and it could involve sequencing hundreds of thousands or even millions of whole genomes to identify the rare variants that explain a substantial portion of the illnesses’ genetic components.
The study reveals only the total amount of rare variants contributing to these common traits — not which ones are important, says Spector. “The next stage is to go and work out which of these rare variants are important for traits or diseases that you want to get a drug for.”