Skip to main content
Premium Trial:

Request an Annual Quote

23andMe Team Sees Genetic, Nongenetic Contributors to COVID-19 Risk in Trans-Ancestry Study

NEW YORK – A team from 23andMe and the GSK Medicines Research Centre has identified genetic and nongenetic factors that appear to impact the risk of infection with SARS-CoV-2 or hospitalization for COVID-19. The effort, announced last spring, used data from more than one million 23andMe participants.

"Given the rapid emergence of COVID-19, preexisting genetic cohorts offer a path to rapid data collection that can address questions surrounding the relationship between host genetics and COVID-19 in a timely fashion," lead authors Janie Shelton and Anjali Shastri, both researchers at 23andMe, and their colleagues wrote.

For a study published in Nature Genetics on Thursday, their team searched for genetic and nongenetic contributors to self-reported SARS-CoV-2 infection, respiratory symptoms, and hospitalization. Specifically, they conducted a genome-wide association study that included genotypes for around 1.05 million 23andMe customers from a range of ancestral backgrounds who had consented to participate in research.

"In under four months, over one million research participants contributed to this study of a new disease via online surveys," the authors noted, though they cautioned that "cases identified in this study were healthy enough to respond to the survey and therefore likely biased toward a healthier case population than otherwise exists."

The group included 15,434 participants diagnosed with COVID-19 and more than 1,100 individuals hospitalized with the disease, the researchers explained. Along with individual genome-wide association studies on participants with European, Latino, or African American ancestry, they brought genetic data from the full cohort together for a trans-ancestry meta-analysis focused on risk factors across ethnic groups.

Similar to past studies, the team saw ties between SARS-CoV-2 infection susceptibility and a chromosome 9 SNP in linkage disequilibrium with an ABO gene variant in individuals from European, Latino, and African American populations. In particular, the type O blood group appeared to offer some protection against SARS-CoV-2 infection compared to the type A, B, or AB blood groups. The data on hand hinted that a similar relationship may exist between COVID-19 severity and blood type, though that association was not statistically significant.

The investigators did see significant ties to respiratory symptoms and COVID-19 disease severity at a gene-heavy chromosome 3 locus called 3p21.31, including a risk allele that appears to be somewhat enriched in individuals of European ancestry.

"We identified a strong association with the ABO gene, which appears to be connected with testing positive for SARS-CoV-2," the authors reported, "and another strong association within a gene-rich locus at chromosome 3p21.31, which appears to be connected with COVID-19 severity."

Even so, the authors warned that "scarcity of testing in the United States during our data collection period likely further obscured the true pictures of SARS-CoV-2 infections, potentially resulting in misestimation of risk factors compared to a sample randomly drawn from the broader population with perfect case and control classification."

The findings line up with those reported by other investigators searching for host genetic clues to COVID-19 disease. At the American Society of Human Genetics virtual meeting last fall, for example, the University of Helsinki's Andrea Ganna shared findings from the COVID-19 Host Genetics Initiative — an international collaboration that included data from 23andMe and other direct-to-consumer genetic testing companies.

That team saw significant associations at nine loci, including the ABO blood locus and the chromosome 3p21.31 locus. The latter site was also implicated in respiratory failure risk by members of a Severe COVID-19 GWAS group that reported their results in the New England Journal of Medicine in October.

Still other studies have highlighted the type I interferon pathway as an apparent contributor to SARS-CoV-2 infection severity, particularly the risk of developing COVID-19 pneumonia, prompting searches for rare variants in type I interferon genes that may impact viral infection risk in pediatric patients.

On the non-genetic side, meanwhile, the new 23andMe analysis pointed to several factors, ranging from advanced age and being male to obesity and pre-existing cardiometabolic conditions, that appear to be involved in risk of hospitalization. But lower socioeconomic status and non-European ancestry also coincided with an increased risk of landing in the hospital with a SARS-CoV-2 infection, highlighting a role for health and social disparities and the importance of looking beyond genetics alone.

"While non-European ancestry was a significant risk factor for hospitalization after adjusting for sociodemographics and preexisting health conditions, we did not find evidence that [the ABO and chromosome 3] primary genetic associations explain risk differences between populations for severe COVID-19 outcomes," the authors reported.

Earlier this year, 23andMe launched a COVID-19 severity prediction tool called the COVID-19 Severity Calculator, which reportedly takes non-genetic risk factors into account.

"Our results strengthen the support for a role in blood type in susceptibility to the pathogen," Shelton said in an email, adding that the severity-related SNP on chromosome 3 "maps to several plausible genes (SLC6A20, CCR9, FYCO1, CXCR6 and XCR1) and work is underway to identify what, specifically, the mechanism of action is."

Shelton noted that the risk calculator is based on a model for hospitalization that was included in the paper.

"In that model, we identified an elevated risk for hospitalization among people with elevated body mass index, pre-existing cardio-metabolic disease, age, and male sex," she wrote. "Adjusting for all of those factors and education and income, we saw an elevated risk of hospitalization for our African American/Black participants, which was unexplained by the data we had in hand."