PROJECT SUMMARY/ABSTRACT Atrial fibrillation (AF) is the most common cardiac arrhythmia and the leading cause of ischemic stroke. Anticoagulants reduce the risk of stroke in AF patients by about two-thirds, but the main safety concern of this long-term oral therapy is the risk of fatal or life-threatening bleeding. Warfarin was once the leading oral anticoagulant in patients with AF, but it is being replaced by Direct Oral Anticoagulants (DOACs). Although AF patients bleed less with DOACs than warfarin, bleeding is still a major problem varying from ~2% to ~40%. Clinical factors predisposing to bleeding are well-known, but they do not fully explain bleeding risk. The best clinical prediction model still does not explain 32% of bleeding events on DOACs. Unlike warfarin, which is monitored by international normalized ratio (INR), there is no routinely used measure for clinical monitoring of DOACs. Therefore, there is a critical need to better understand the risk factors for bleeding from DOACs. Since drug responses are highly heritable, the central hypothesis of this research is that genetics can help explain bleeding risk in AF patients on DOACs. Therefore, the overall goal of this research is to identify genetic variants associated with bleeding risk from DOACs. Previous candidate gene studies associated genetic variants with changes in plasma concentration of DOACs, which may increase bleeding risk in AF patients. However, there are only a few candidate gene studies assessing bleeding as an outcome with very small sample sizes (typically n <400) and inconclusive results. Therefore, the objective of Aim 1 is to overcome those limitations by performing candidate gene association analysis of a large clinical & genomic dataset with bleeding outcomes (n=2,470 AF patients on DOACs in the Michigan Genomics Initiative [MGI]). Another limitation of the previous studies was the reliance on the candidate gene approach, which can miss unsuspected genes. A genome-wide association study (GWAS) could discover novel variants associated with the risk of bleeding from DOACs and enable the derivation of the first polygenic score for predicting this risk, which are the objective of Aim 2. Both approaches will be carried out to compensate for the limitations of each method. The overall approach is to analyze an existing clinical & genomic dataset (MGI), which integrates whole genome array data (post quality control & imputation) and complete access to electronic medical records (EMR). The primary outcome is a composite of major and clinically relevant non-major (CRNM) bleeding. The secondary outcomes are major and clinically relevant non- major bleeding. This research is feasible by leveraging an existing dataset, the expertise of all sponsors, resources at the University of Michigan, and established methods. Also, this research is highly relevant because it addresses an important clinical knowledge gap by determining the association of genetic variants with bleedin...