The study

The human genome, 5 million variants: the risk of over-interpreting DNA

Finding a genetic variation does not necessarily mean predicting a disease, and the risk of attaching excessive clinical significance to uncertain results increases as the amount of available information grows

by Francesca Cerati

22 October 2026

Aggiungi Il Sole 24 Ore
ai preferiti su Google

3' min read

Translated by AI

Versione italiana

3' min read

Translated by AI

Versione italiana

As genetic sequencing becomes increasingly accessible, the likelihood of identifying variants associated with diseases – even in healthy people – is growing. But a genetic predisposition is not a diagnosis: a study of nearly a million genomes sets out what to expect from each of our DNA and provides new guidelines for interpreting the test results.

Sequencing DNA to diagnose a disease, identify a predisposition or assess the risk of developing a condition in the future: with the growing availability of genetic testing, these possibilities are expanding. But the more our ability to identify variants associated with diseases increases, the more important it becomes to understand what they actually mean. Because finding a genetic alteration does not necessarily mean predicting a disease, and the risk of attributing excessive clinical significance to uncertain results grows in line with the amount of information available.

A study by the University of Exeter, published in *Genome Medicine*, has taken stock of this issue; it analysed data from nearly one million people involved in two large biobanks, the UK Biobank in the UK and the US-based ‘All of Us’ project. The aim is to define what constitutes an ‘average’ human genome, providing doctors and researchers with a benchmark for distinguishing normal genetic variability from alterations that may be relevant to health.

Five million variants per individual

On average, each person has around five million genetic variants, that is, positions in the DNA where the sequence differs from the reference sequence. Around 10,000 of these affect gene sequences and can alter their function, with a wide range of consequences: from traits such as lactose intolerance to differences in response to medicines, right through to alterations associated with diseases. Many variants, however, have no clinically significant effects.

The analysis also shows that each individual generally carries between one and three rare variants already associated with serious genetic diseases. This finding does not amount to a diagnosis: some variants, for example, may indicate carrier status without leading to the development of the condition. In addition to these, there are around 50 rare variants whose clinical significance remains uncertain, as current knowledge does not allow their effects to be determined with certainty.

Genetic risk is not a diagnosis

The researchers also examined polygenic risk scores, which are tools that combine the effects of numerous variants to estimate a person’s predisposition to common diseases. Looking at 28 conditions, including diabetes, cardiovascular disease and schizophrenia, they found that most people would fall into the highest genetic risk category for around three diseases.

The findings should be interpreted with caution. The study did not assess the participants’ health status and therefore does not prove that people classified as genetically predisposed have developed or will develop these conditions. Genetic risk is only one component of a broader picture, which also involves age, environment, lifestyle, family history and other clinical factors.

“Variability is the norm,” emphasises Caroline Wright, a lecturer at the University of Exeter and senior author of the study. With the increasing use of whole-genome sequencing and predictive screening, understanding how many variants are normally present in each individual is essential to avoid misinterpretation of the results.

The problem is not merely to distinguish pathogenic variants from harmless ones, but also to understand what they mean for that specific individual. Symptoms, family history and other risk factors must all be taken into account in the clinical assessment, without turning a genetic result into an inevitable prediction.

New mutations and the father’s age

The study also examined de novo mutations – those that appear for the first time in an individual’s genome and are not detectable in that of their parents. Among those for whom family data were available, each individual had, on average, around 63 new variants of this type. Their number increased with the father’s age at conception: approximately 1.2 additional mutations for every year of the father’s age.

This is a biologically significant association, largely linked to the cell divisions that occur during sperm production. This does not, however, mean that every new mutation is harmful: most variants do not necessarily lead to disease, and this information alone does not allow us to predict the outcome of an individual pregnancy.

The authors also acknowledge a significant limitation: the two biobanks do not provide a balanced representation of all the world’s populations, with a significant proportion of individuals of Western European origin. Increasing the diversity of the data will be necessary to improve the interpretation of variants and the reliability of polygenic scores across different populations.

The value of this work, therefore, does not lie in identifying an ‘ideal’ genome against which to compare all others, but in better defining the variability that characterises the human species. This is an important step forward for genomic medicine, provided that the growing ability to sequence DNA is matched by the ability to interpret it correctly.