Human genomics research continues to have a major bias towards people of European ancestry, scientists say – and that could have damaging consequences in terms of how research is interpreted and followed up.

The statistics tell their own story: as of 2018, individuals included in genome-wide association studies (GWAS) were 78 percent European, 10 percent Asian, 2 percent African, 1 percent Hispanic, and less than 1 percent all other ethnic groups.

As these GWAS studies are used to predict disease risk, develop medical treatments, and plan further research and study, the risk is that we're not seeing the whole picture of the human population – even if the individual studies themselves are scientifically sound.

"Leaving entire populations out of human genetic studies is both scientifically damaging and unfair," says one of the researchers, evolutionary geneticist Sarah Tishkoff from the University of Pennsylvania.

"We may be missing genetic variants that play an important role in health and disease across ethnically diverse populations, which may have deleterious consequences in terms of disease prevention and treatment."

Tishkoff and her colleagues looked at the thousands of publications listed in the GWAS Catalog to come up with their figures, as well as analysing genetic risk in particular health issues like kidney disease and schizophrenia.

Applying findings on genetic risk taken from Europeans might not necessarily work in non-Europeans, the team argues – because of variations passed down through evolutionary history, as humans originated from and spread to different areas across hundreds of thousands of years.

Some diseases are linked to a single gene variant, but others are associated with many different genes, as well as environmental factors – it's here that the lack of a wide, unbiased sample set becomes a real problem. 

"The lack of diversity in human genomics studies is likely to exacerbate health inequalities," says one of the team behind the new research, Scott M. Williams from the Case Western Reserve University School of Medicine in Ohio.

"For example, approaches are being developed to predict a person's risk of diseases such as Alzheimer's disease, heart disease, or diabetes based on their status for multiple genes. But such calculations developed based on evidence from primarily European populations may not apply to people of other ethnic backgrounds."

The researchers give the example of cystic fibrosis, about six times more common in those of European descent than African descent. The most common causative allele in the first group accounts for 70 percent of cases, but only 29 percent of cases in the second group.

Specific genetic mutations might be happening in populations we haven't studied enough, the team suggests, and together with the effects of genetic drift as populations separate, it means the end results of GWAS research might not be as precise as we would like.

The solution put forward by the researchers is to use comprehensive biobanks for future studies, wherever possible: biobanks with ethnically diverse individuals that can be linked to extensive health records. That should result in better healthcare for everyone.

"These initiatives will require the political will to improve funding and infrastructure for studying genomic and phenotypic diversity in global populations," says one of the researchers, Giorgio Sirugo from the University of Pennsylvania.

"The future success of genomic and precision medicine depends on it."

The research has been published in Cell.