Skip to Main Content

On March 16, 2020, five days after the World Health Organization declared Covid-19 a global pandemic, Andrea Ganna, a geneticist at the Institute for Molecular Medicine Finland, took to Twitter to make an announcement: “We are launching the ‘COVID-19 host genetics initiative,’” he wrote. He asked other scientists to join him and institute director Mark Daly in probing the world’s stores of human DNA to help answer a complicated but pressing question: Why do some unlucky people infected by the coronavirus end up gasping for air in an intensive care unit, while many others catch and spread the disease without having so much as a cough?

Ganna wasn’t quite prepared for how many would take him up on the offer. Over the last 15 months, more than 3,300 researchers from 25 countries have poured data from millions of people, including more than 125,000 Covid-19 patients, into the initiative, making it one of the largest gene-hunting missions in history.


The international effort has revealed that an individual’s genetic inheritance can indeed influence their risk of infection and the severity of disease.

On Thursday, Ganna and a long list of co-authors reported in Nature that more than a dozen parts of the human genome were linked with either enhanced susceptibility to infection with SARS-CoV-2 or severe Covid-19. The research won’t change current treatment decisions for patients, but these genetic clues may point to existing drugs that could be repurposed to help the worst-off among them. Outside the heavily vaccinated U.S., Covid-19 is still raging, driven by the arrival of the dangerous new Delta variant, and effective treatments are still limited to expensive monoclonal antibody cocktails, the steroid dexamethasone, and two inflammation-easing arthritis drugs.

“We have quite clear and compelling genetic signals that point to some expected, but also some new biological processes,” Ganna told reporters at a press briefing Wednesday. “Some of these are specific to Covid-19 and not necessarily shared with other infectious diseases.”


For most infectious diseases, common genetic variants — a few non-life-threatening DNA letter changes here and there throughout the genome — tend to alter the course of infection only slightly. Age, occupational exposure, and access to quality health care matter far more. That makes detecting any genetic influence difficult. But SARS-CoV-2 emerged at a time when scientists around the world had already begun to stockpile DNA data from millions of research participants. When some of those people became infected with the coronavirus, it just became a matter of using statistical techniques to look for patterns.

One of these methods, a genome-wide association study, or GWAS, involves sorting people into different groups — Did they catch Covid-19? Did they get sick enough to go to the hospital? Did they die? — and then scanning their DNA to see if certain single DNA letter variations show up more often among people in each group.

Using this standard technique, previous DNA-mining efforts unearthed a handful of genes that seemed to play a role in swaying the course of a coronavirus infection, including the genes that code for a person’s blood type, a gene implicated in fending off influenza, and a stretch of chromosome 3 containing genes involved in ramping up host immune defenses. The Nature study confirms those previous discoveries and adds six new ones.

These include changes to the regions near four genes, of which the most interesting is one called FOXP4. The same changes that showed up in Covid-19 patients with more severe forms of the disease have also been linked to lung cancer. In both cases, the changes led to increased expression of FOXP4.

“This gives us a plausible hypothesis by which the repurposing of drugs, if a FOXP4 inhibitor was available, for example, could plausibly be taken forward,” Daly said at the briefing. FOXP4 is believed to be involved in T cell development, and its role in a variety of cancers is just beginning to be investigated. Notably, this variant occurs more frequently in individuals with East Asian ancestry, and it was only discovered when cohorts from that part of the world were included in the analysis.

Other experts applauded the effort for its scale of collaboration and speed at generating potential leads. But they cautioned that there was much further work to be done before any drug repurposing efforts inspired by the GWAS results could advance to clinical trials.

“It’s exciting to have this list, but it’s just a starting point,” said Nevan Krogan, director of the Quantitative Biosciences Institute at the University of California, San Francisco, who was not involved in the new study. “Experiments still need to be done to figure out the biology behind each variant.”

Krogan’s group at UCSF launched its own international collaboration at the start of the pandemic aimed at speeding up the search for potential drug candidates, but using a different approach to map out all the protein interactions between SARS-CoV-2 and its human hosts. Based partially on work they published in Nature last spring, 26 clinical trials are now underway. One of these, a cancer drug called plitidepsin, is currently in Phase 3 study in patients with moderate Covid-19.

By comparison, Daly said he was not aware of any Covid-19 drug trials underway that targeted genes uncovered by their analyses. “But, these genetic results are still relatively new,” he said.

The COVID-19 Host Genetics Initiative has been posting periodic updates as new data become available. The results reported in Nature represent information available as of January, constituting data from 50,000 Covid-19 patients and 2 million non-infected controls. The group has since added a new update to its website with data from 125,000 patients. The latest analysis adds an additional 10 genetic variants, including one upstream of the gene for ACE2, the receptor SARS-CoV-2 uses to infect human cells.

This variant was also reported by a group from the Regeneron Genetics Center in a preprint last month. Unlike many of the previously discovered genetic changes, this one was protective — corresponding to a 40% reduction in risk of severe Covid-19. And it was located on a sex chromosome, in this case the X chromosome, which many GWAS analyses leave out.

Intrigued, the Regeneron researchers turned to an RNA expression dataset, and found that the genetic variant lowered levels of ACE2. The effect was also much stronger in males than females. “We have this thread showing that a decrease in ACE2 expression is protective against SARS-CoV-2 infections,” said lead author Julie Horowitz. “That’s really useful when thinking about ACE2 as a therapeutic target because now we know that blocking it through antibodies or other mechanisms might prevent infection or severe disease.”

One of her colleagues combined the effects of this variant and all the others researchers had found into an assessment of how an individuals’ genetic makeup might influence the course of Covid-19, known as a polygenic risk score. For people with European ancestry, it predicts that one’s DNA could, at most, make you 1.4 times more likely to be hospitalized from the coronavirus. That’s not much of a jump, but it’s similar to what companies like 23andMe have found with its polygenic risk score for type 2 diabetes. (The results were similar but less reliable for individuals with non-European ancestry, who only made up about 20% of the GWAS dataset.)

While at least one company is marketing a polygenic risk test for Covid in the U.S. based on genetic signals found by the COVID-19 Host Genetics Initiative, scientists leading that effort said the data are still too murky for it to be useful to doctors guiding decisions for patients. “Clearly there is a role for genetics in [predicting] COVID severity, but I would not think it’s overwhelming,” said Ganna. “It’s one of many risk factors.”

One of those that Ganna and his colleagues have yet to look at is how different strains of SARS-CoV-2 interact with someone’s genetic background to impact the severity of disease. The datasets they had access to did not include viral sequence data, so they have thus far been limited to looking at just the host side of the equation. But it’s an area of inquiry they hope their network will soon be in a position to tackle.

That’s exactly where the science should be going next, said Krogan, and not just for Covid-19. “We should be looking at both the virus and the host because an infection is an interaction between its genetic material and ours,” he said. “Where this is all heading is something that looks like precision medicine for infectious disease, where sequencing tells us both the genetic makeup of the patient and the virus infecting them, and that tells you which treatment to get. We’re not there now. But in the future you could envision we have the information to do it.”

Create a display name to comment

This name will appear with your comment

There was an error saving your display name. Please check and try again.