Main author: Þórarinn Jónmundsson
Institution or Company: School of Health Sciences University of Iceland 101 Reykjavik Iceland
Co-Authors, Institution or Company:
Thor Aspelund, Icelandic Heart Association Holtasmari 1 201 Kopavogur Iceland, Faculty of Medicine University of Iceland 101 Reykjavik Iceland. Valur Emilsson, Icelandic Heart Association Holtasmari 1 201 Kopavogur Iceland, Faculty of Medicine University of Iceland 101 Reykjavik Iceland. Elías Freyr Guðmundsson, Icelandic Heart Association Holtasmari 1 201 Kopavogur Iceland. Valborg Guðmundsdóttir, Icelandic Heart Association Holtasmari 1 201 Kopavogur, Iceland, Faculty of Medicine University of Iceland 101 Reykjavik Iceland. Vilmundur Guðnason, Icelandic Heart Association Holtasmari 1 201 Kopavogur Iceland, Faculty of Medicine University of Iceland 101 Reykjavik Iceland.
Introduction: Myocardial infarction (MI) is caused by blocked blood flow in the atherosclerotic coronary arteries. The discovery of causal relationships between specific proteins and MI raises the prospect of drug development. However, conducting randomized controlled trials to prove causation is difficult. To address this problem a statistical method, Mendelian randomization, has been developed using single nucleotide polymorphisms (SNP) as proxies for the randomization. This study applies an MR approach on serum proteins in the AGES cohort to prioritize causal candidates for MI.
Methods: A two sample MR study was conducted. Genetic predictors (SNPs) for 4137 human serum proteins within a cis-window of 500 kilobases (kb) up- or downstream of the respective protein-encoding gene were obtained from population based AGES-Reykjavik which consists of 5368 participants. SNPs were filtered to only include independent genetic instruments that were not in linkage disequilibrium (r2 < 0.2) or outside of a 500 kb window. To associate the predictors with MI summary statistics from the CARDIOMGRAMplusC4D 1000 Genomes based GWAS were used. The causal effect of the SNPs were estimated with the inverse variance weighted and Wald ratio estimators.
Results: For MI, 78 unique serum proteins were supported as causal after adjusting the false discovery rate. Among these were proteins with well-established connections to MI such as APOE, APOA5 and PCSK9.
Conclusion: The findings are preliminary, but promising, and corroborate previous findings. The pleiotropic effects of the instruments, as well as the course of causality, will be investigated further in the future for validation.
