Survey of nearly 1 million genomes finds disease-linked DNA variants in everyone

Everyone's genome, however healthy they are, contains a handful of mutations already linked to serious disease and places them in a high genetic-risk category for several common conditions, according to the largest survey yet of what a typical human genome actually looks like.

The study, led by the University of Exeter and published Thursday in Genome Medicine, analyzed whole-genome sequences from nearly one million people drawn from the UK Biobank and the United States' All of Us research program. It is being presented this week at the American Society of Human Genetics annual meeting in Montreal.

The researchers set out to answer a basic but previously unquantified question: what does an ordinary genome contain, once the enormous scale of modern biobanks makes it possible to establish a statistical baseline. Their answer complicates a growing assumption in medicine and consumer genetics — that finding a disease-linked variant in someone's DNA is itself meaningful news.

What a million genomes show

Each person's genome carries roughly five million variants, places where an individual's DNA differs from the reference sequence. Of those, about 10,000 alter the sequence of a gene in ways that can affect anything from lactose tolerance to how a person metabolizes certain drugs, though the Exeter team found most such changes are biologically neutral.

More strikingly, every person in the dataset carried between one and three rare variants that had previously been linked to serious genetic disease, plus roughly 50 additional rare variants whose clinical significance remains uncertain. The researchers did not assess participants' actual health status, so many of the people carrying these variants may never develop any related illness.

The team also calculated polygenic risk scores — composite measures that combine many common variants to estimate susceptibility to disease — for 28 common conditions, including type 2 diabetes, cardiovascular disease and schizophrenia. On average, most people fell into the high-risk category for around three of those 28 conditions. A polygenic risk score, as the National Human Genome Research Institute defines it, estimates relative rather than absolute risk and does not account for environment or lifestyle.

A separate finding concerned new mutations that arise spontaneously rather than being inherited. Among participants who had been sequenced alongside both parents, each person carried around 63 of these "de novo" variants not present in either parent's genome. Most originated in the father's sperm, and the number rose by about 1.2 variants for every additional year of the father's age at conception.

A problem for genetic interpretation

The findings bear directly on a difficulty clinicians and genetic counselors already grapple with: distinguishing a variant that causes disease from one that merely exists. As whole-genome sequencing has become routine for diagnosing rare genetic disorders, and as direct-to-consumer tests and population screening programs expand, the number of "incidental" findings — disease-linked variants turned up in people with no symptoms — has grown accordingly.

Because the Exeter team's figures describe what is actually normal across a population of nearly a million people, they give clinicians a numerical baseline against which to judge how alarmed to be about any single result. A finding that would once have looked like a smoking gun — a rare variant tied to a serious disease — turns out to be something close to universal.

"Whole genome sequencing has become such an important tool in diagnosing genetic disease, and we're increasingly moving towards genetic screening for predicting disease. Yet most people are unaware of just how much genetic variation is present in every genome. Actually, variation is the norm," said Professor Caroline Wright, the study's senior author and a professor of genomic medicine at the University of Exeter. "Importantly, we found that everyone carries a small number of genetic variants linked to disease, although many of these will not actually cause disease in that individual. It is therefore crucial to consider context when interpreting findings from genetic testing, including symptoms or family history of disease as well as other clinical and environmental risk factors."

Gaps in the data, and what comes next

Wright's team is explicit about the study's central limitation: both the UK Biobank and All of Us cohorts, despite their size, remain skewed toward participants of Western European ancestry. The researchers say the pattern of genetic variation they describe may not generalize fully to other populations, and they call for equivalent sequencing efforts in more genetically diverse groups to refine the baseline.

The practical upshot, according to the Exeter release accompanying the paper, is aimed less at the public than at the professionals who interpret genetic results: physicians ordering sequencing for a patient with an unexplained illness, genetic counselors discussing a biobank return-of-results program, and health systems designing population-wide screening. The study's authors frame their data as a reference tool for exactly those judgment calls, intended to reduce both unwarranted alarm and false reassurance.

The work adds to a broader trend in genomics, in which the falling cost of sequencing and the growth of biobanks holding data on hundreds of thousands or millions of participants have shifted research away from studying rare diseases in isolation and toward characterizing ordinary genetic variation at population scale. The Exeter paper is a product of funding from the Medical Research Council, Wellcome and the National Institute for Health and Care Research's Biomedical Research Centre in Exeter, and arrives as sequencing-based screening programs are being piloted or expanded in several health systems, raising the stakes for getting the interpretive baseline right.