Exome-Wide Association Study Reveals Host Genetic Variants Likely Associated with the Severity of COVID-19 in Patients of European Ancestry

Abstract

Host genetic variability plays a pivotal role in modulating COVID-19 clinical outcomes. Despite the functional relevance of protein-coding regions, rare variants located here are less likely to completely explain the considerable numbers of acutely affected COVID-19 patients worldwide. Using an exome-wide association approach, with individuals of European descent, we sought to identify common coding variants linked with variation in COVID-19 severity. Herein, cohort 1 compared non-hospitalized (controls) and hospitalized (cases) individuals, and in cohort 2, hospitalized subjects requiring respiratory support (cases) were compared to those not requiring it (controls). 229 and 111 variants differed significantly between cases and controls in cohorts 1 and 2, respectively. This included FBXO34, CNTN2, and TMCC2 previously linked with COVID-19 severity using association studies. Overall, we report SNPs in 26 known and 12 novel candidate genes with strong molecular evidence implicating them in the pathophysiology of life-threatening COVID-19 and post-recovery sequelae. Of these few notable known genes include, HLA-DQB1, AHSG, ALOX5AP, MUC5AC, SMPD1, SPG7, SPEG,GAS6, and SERPINA12. These results enhance our understanding of the pathomechanisms underlying the COVID-19 clinical spectrum and may be exploited to prioritize biomarkers for predicting disease severity, as well as to improve treatment strategies in individuals of European ancestry.

Keywords: COVID-19 host genetics; common genetic variants; exome-wide association study for COVID-19 patients; genetic variation in COVID-19 patients.

Figure 1

Principal Component Analysis (PCA) of COVID-19 patient genomes. PCA plot showing genetic differentiation among COVID-19 patient genomes. COVID-19 patients (NH, HWRS, HWOS, HWCB, HI, and D) were designated in various shades of grey and dead individuals were designated with black. African, South Asian, European, Latin American, and East Asian populations were designated with different shades of yellow, red, blue, pink, and green, respectively. We selected COVID-19 patients that cluster with European genomes (PC1 ranging from −0.0050 to 0.0050 and PC2 ranging from −0.0100 to 0) for downstream analysis. PCA was performed in PLINK v1.9. The PC1 and PC2 were plotted in RStudio v1.4.1717.

Figure 2

Manhattan Plots summarizing Exome-wide Association Study results. X-axis represents chromosomes (chr 1 to chr Y). SNPs present in the chromosomes are designated with dots. Negative log-transformed (−log10) covariate adjusted p-values are plotted in the Y-axis. The SNPs with p-value < 0.00001 are indicated with the blue line, and those with p-value < 0.0000001 are indicated with the red line. (A). cohort 1. Genomes of non-hospitalized COVID-19 patients (N = 493) were compared against the hospitalized patients (N = 2199). (B). cohort 2. Genomes of COVID-19 patients with respiratory support (N = 1877) were compared against those without respiratory support (N = 815).

Figure 3

FoamTree representing various Reactome pathways associated with the significant SNPs identified in (A). cohort 1 and (B). cohort 2. Pathway map was generated using SNPnexus web-based server. Pathways associated with the submitted dataset are highlighted in various shades of yellow. The gray entities represent the pathways that are represented in the query dataset but absent in the submitted dataset.