Observational causality from -omics

Some human traits like disease are heritable, which means that they run in families. This indicates that there must be something on the DNA that affects an individual’s susceptibility to developing a trait. In the last 15 years, scientists from around the world have been very successful in mapping the locations on the DNA that are associated to traits like disease, finding thousands of loci, and hundreds of DNA locations per trait, making them truly complex traits. So, we have a very good understanding about which locations on the DNA are important for developing complex traits like disease. Unfortunately, it’s still unclear how these locations on the DNA affect an individual’s trait.
In this thesis I investigate how we can best understand the DNA locations that affect trait susceptibility and in doing so, identify the causes for human traits like disease. One important technique that we have used to test for finding these causal relationships is called Mendelian randomization. Mendelian randomization identifies naturally occurring experiments that have happened in observational data. In principle, Mendelian randomization can conclude the same things from observational data as from an experimental study. So called `observational causality` has many benefits as it’s cheaper than an experiment, and is less burdensome on the subjects, as they are not subjected to any intervention.
The causes that I’m interested in are so called `-omics` traits. -omics traits are molecular measurements that are usually strongly regulated by the DNA. This strong DNA regulation makes -omics traits interesting candidates to understand the mechanism behind the genetic loci of other traits.
In this thesis we have investigated gene expression, protein levels and microbiome measurements as our -omics traits of interest for a wide variety of traits including celiac disease and LDL-cholesterol levels.