Using self-reported survey data, researchers have defined eight phenotypes associated with coronavirus disease 2019 (COVID-19) outcomes. Analysis of the data revealed possible protective genetic associations.
Several genome-wide association studies (GWAS) of COVID-19 have identified genetic associations that suggest susceptibility and risk of infection by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may be hereditary.
Most of these studies were performed on hospital patients, however, which can lead to a bias toward severe disease outcomes.
To circumvent this and identify traits, or phenotypes, that may predict COVID-19 outcomes, researchers from AncestryDNA used a survey of their customers that assessed patients’ exposure, risk factors, symptoms and demographics, most of whom had mild disease.
The team collected self-reported data using a survey platform available to AncestryDNA customers and reported their results in the medRxiv* preprint server.
Using more than 700,000 survey results collected between April and August 2020, the team defined eight phenotypes to study further. Four of these were similar to susceptibility phenotypes used in previous studies and four were new.
The team hypothesized that traits focusing on mild symptoms may better help identify protective genetic associations.
Genetic associations for mild COVID-19
To help identify phenotypes for infection susceptibility, the authors defined two new phenotypes that focused on respondents who were exposed to the virus within households, comparing people who were negative but had household members who tested positive with a control population.
The other two phenotypes compared symptomatic to mild or asymptomatic patients and asymptomatic and severely ill patients.
Using 13 independent single nucleotide polymorphisms (SNPs) that showed genome-wide significance in two recent studies, the team investigated their association with the eight phenotypes the team identified.
The ancestry of the population analyzed included 73 percent European, 6 percent Amerindian, and 3 percent African-European.
10 of the 13 SNPs were found in at least one of the eight phenotypes, indicating that self-reported outcomes are similar with associations found in clinical phenotypes.
They found three independent locus replication signals on chromosome 3 near the genes LZTFL1 and SLC6A20, which are related to immunity.
The previously identified SNP rs35081325 was associated with the severity of phenotypes, suggesting increased risk of infection severity, while rs73062389 was associated with the susceptibility phenotypes and not with any of the severity cluster phenotypes.
They found that the third SNP rs2531743 was associated with two phenotypes and indicated a protective effect.
Thus, the signals in chromosome 3 are associated with different sets of phenotypes. This suggests that variations in this region alter many aspects of disease severity and susceptibility.
Studies have also shown associations near the ABO gene, which determines blood type. The ABO SNP rs505922 was seen in all the four susceptibility phenotypes and one severity phenotype.
Protective genetic associations
The authors next performed a discovery GWAS for all eight phenotypes to identify new phenotype-locus associations. Although no phenotypes reached a high level of association, the team looked at associations that reached suggestive levels.
It was found that minor alleles associated with three new phenotypes defined in the study were almost always associated with a protective effect. This is in contrast to other studies in which the minor alleles were associated with risk.
Phenotype definitions that are based on mild or absence of symptoms, even with strong exposure, may therefore be better at identifying protective genetic associations.
Replication rates among the 297 phenotype-locus association pairs were also found to be generally low and there was a low correlation among the three populations used for their analysis. It is possible that this is because the replication cohorts had low power or the suggestive associations were false positive.
However, two phenotypes, one associated with severity and another with susceptibility, showed a positive correlation, suggesting that they might give reproducible associations with a larger sample size.
The team found that the association of the SNP rs55673936 with whether or not a patient was hospitalized was reproducible in two independent populations. This SNP on chromosome 11 in the gene GALNT18 was previously reported to be associated with increased response to Tocilizumab, an IL-6 blocking monoclonal antibody.
The study provides a complementary analysis to studies focusing on severe disease, which is promising for finding new genetic associations, in particular those that provide protection against the virus.
medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.