Many current drugs for stroke act by targeting circulating molecules, yet these have not been exhaustively evaluated for therapeutic potential. A central challenge is that while many molecules correlate with stroke risk, only a subset cause stroke. To disentangle causality from association, a statistical genetics framework called “Mendelian Randomization” can be used by integrating genetic, biomarker, and phenotypic information. In Study 1, we screened 653 circulating proteins using this technique and found evidence supporting causal roles for seven proteins, two of which (SCARA5 and TNFSF12) were not previously implicated in stroke pathogenesis. We also characterized potential side-effects of targeting these molecules for stroke prevention and did not identify any adverse effects for SCARA5. The remaining two studies focused on investigating the role of an emerging marker of mitochondrial activity, leukocyte mitochondrial DNA copy number (mtDNA-CN). Mitochondria have long been known to play a protective role in stroke recovery; however, a mitochondrial basis for stroke protection has not been extensively studied in humans. In Study 2, we first sought to better understand the genetic basis of mtDNA-CN in a series of genetic association studies involving 395,781 UK residents. We identified 71 loci which represents a 40% increase in our knowledge. In Study 3, epidemiological analyses of 3,498 acute stroke demonstrated that low mtDNA-CN was associated with higher risk of subsequent mortality and worse functional outcome 1-month after stroke. Furthermore, Mendelian Randomization analyses corroborated a causative relationship for the first time, implying that interventions that increase mtDNA-CN levels in stroke patients may represent a novel strategy for mitigating post-stroke complications. Ultimately, this work uncovered several novel therapeutic leads for preventing stroke onset and ameliorating its progression. Future investigations are necessary to better understand the underlying biological mechanisms connecting these molecules to stroke and to further interrogate their validity as potential drug targets.