Exploiting shared omics architecture and causal association between complex traits and COVID-19

Background Understanding the shared biological and environmental correlations between COVID-19 and relevant complex traits may provide insights to mechanistic pathways and shared markers. Methods In this study, we employed bivariate Genomic-based Restricted Maximum Likelihood (GREML) to assess correlations between COVID-19 and 35 relevant complex traits, followed by Mendelian randomization (MR) to investigate potential causal links. To do that, we used phenotypic and omics data (genomic, transcriptomic, proteomic, metabolomic, and exposomic) from up to 107,857 participants from the UK Biobank. Results No genetic or transcriptomic correlations were found between COVID-19 and the traits. However, proteomic analysis revealed significant positive correlations with cardiometabolic traits (e.g., from r = 0.28 for chronic renal failure to r = 0.43 for heart disease) and pneumonia (r = 0.22). It also showed negative correlation with peptic ulcer (r = -0.26) and biomarkers (cystatinc-c: r = -0.19 and urea: r = -0.23). Metabolomic correlations identified strong associations with hormone-sensitive cancer (r = 0.72) and obesity-related cancer (r = 0.60), while negative correlations were observed for HDL cholesterol ( r= -0.34) and vitamin D (r = -0.32). Exposomic correlations mirrored proteomic findings with significant links with cardiometabolic traits such as heart diseases (r = 0.52) and stroke (r = 0.56). MR analysis indicated that none of the examined traits exhibited causal effects on COVID-19, suggesting that the observed omics correlations resulted from horizontal pleotropy, where omics factors independently influence both COVID-19 and complex traits. Conclusions Our study has demonstrated a shared omics architecture at proximal layers between COVID-19 and relevant traits. However, MR analysis showed no causal effects of these traits on COVID-19, suggesting distinct causal mechanisms from the omics layers.