Approved Research

Determinants and predictors of multiple chronic diseases and traits: identifying shared and unique molecular pathways.

Lay summary

Complex chronic human disorders such as high blood pressure involve interactions among multiple diseases, risk factors and molecular components that need to be jointly taken into account to (1) gain a better understanding of the biological pathways (2) achieve more accurate predictions of individual risk. Genetic variation within the human genome can be used as an instrument to help decipher shared and unique molecular pathways. For example, genetic variants robustly associated to complex traits and proteomics generate insights on how the genotype could be linked to the phenotype through specific proteins. For the next three years, we aim to dissect differential molecular features across multiple GWAS loci and across multiple diseases. Initially, several separate analyses will independently focus on each disease and condition. We will produce robust predictors of molecular traits across omics studies (e.g. transcriptomic, epigenomic, proteomic and metabolomic markers) and imaging measured together with genetics in the UK biobank population. We will ask, which local and distal genetic effects of molecular intermediates are most predictive of disease? Classical population genetics methods and machine learning (ML) classifiers will be used to provide the best predictors. In a second stage, we will compare findings across multiple diseases and identify pathways affecting the unique or multiple complex diseases. The information from UK biobank allows to identify which components that link to each other at the cellular level show higher comorbidity in the population. We will ask, which genetic effects are more similar/dissimilar across diseases, and determine whether differences in the comorbidity patterns indicate differences in genetic background. We will integrate with drug banks to allow validation with respect to the clinical known drug effects and side-effects. We aim to reproduce the genetic effects on multiple phenotypes, and distinguish the downstream biomarkers known to be affected by the shared pathways. This will allow us to gain a better understanding of treatment of a wide range of illnesses and of multimorbidity. Incorporating information of shared predictors and pathways has the potential to increase risk prediction accuracy; we will explore the benefit to healthcare in treating multiple conditions simultaneously.