Genetic architecture of glycomic and lipidomic phenotypes in isolated populations

Abstract

Understanding how genetics contributes to the variation of complex traits and diseases is one of the key objectives of current medical studies. To date, a large portion of this genetic variation still needs to be identified, especially considering the contribution of low-frequency and rare variants. Omics data, such as proteomics and metabolomics, are extensively employed in genetic association studies as ‘proxies’ for traits or diseases of interest. They are regarded as “intermediate” traits: measurable manifestations of more complex phenotypes (e.g., cholesterol levels for cardiovascular diseases), often more strongly associated with genetic variation and having a clearer functional link than the endpoint or disease of interest. Accordingly, the genetics of omics have the potential to offer insights into relevant biological mechanisms and pathways and point to new drug targets or diagnostic biomarkers. The main goal of this thesis is to expand the current knowledge about the genetic architecture of protein glycomics and bile acid lipidomics, two under-studied omic traits, but which are involved in several common diseases. First, in Chapter 2 I compared genetic regulation of glycosylation of two different proteins, transferrin and immunoglobulin G (IgG). By performing a genome-wide association study (GWAS) of ~2000 European samples, I identified 10 loci significantly associated with transferrin glycosylation, 9 of which were previously not reported as being related with the glycosylation of this protein. Comparing these with IgG glycosylation-associated genes, I noted both protein-specific and shared associations. These shared associations are likely regulated by different causal variants, suggesting that glycosylation of transferrin and IgG is genetically regulated by both shared and protein-specific mechanisms. Next, in Chapter 3 I investigated the effect of rare (MAF<5%) predicted loss-of-function (pLOF) and missense variants on the glycome of transferrin and IgG in ~3000 samples of European ancestry. Using multiple gene-based aggregation tests, I identified 16 significant gene-based associations for transferrin and 32 for IgG glycan traits,located in 6 genes already known to have a biological link to protein glycosylation but also in 2 genes which have not been previously reported. Finally, in Chapter 4 I applied a similar approach to bile acid lipidomics, exploring the genetic contribution of both common and rare variants. Despite more than double the sample size (N = ~5000) compared to protein glycomics analysis, I identified only 2 loci, near the SLCO1B1 and PRKG1 genes, significantly associated with bile acid traits., for which I noted a sex-specific effect. Further, I found 3 rare variant gene-based associations, in genes not previously reported as associated with bile acid levels. While the biological mechanisms linking these genes to levels of bile acid is not immediately clear, there is evidence in the literature of their involvement in bile acid synthesis and secretion and in liver diseases. In summary, in my thesis I describe the genetic architecture of the protein glycome and the bile acid lipidome: the former has a higher genetic component, while the latter is largely influenced by environmental factors (e.g., sex, diet, gut flora). Despite the limited sample size, we were able to describe rare variant associations, demonstrating that isolated populations represent a useful strategy to increase statistical power. However, additional statistical power is needed to identify the possible effect of protein glycome and bile acid lipidome on complex disease. A clearer understanding of the genetic architecture of omics traits is crucial to develop informed disease screening tests, to improve disease diagnosis and prognosis, and finally to design innovative and more customised treatment strategies to enhance human health.