| The precise pathophysiology of ischaemic stroke is unclear, and a greater understanding of the different mechanisms that underlie large-artery, cardioembolic and lacunar ischaemic stroke subtypes would enable the development of more-effective, subtype-specific therapies. Genome-wide association studies (GWASs) are identifying novel genetic variants that associate with the risk of stroke. These associations provide insight into the pathophysiological mechanisms, and present opportunities for novel therapeutic approaches. In this Review, we summarize the genetic variants that have been linked to ischaemic stroke in GWASs to date and discuss the implications of these associations for both our understanding and treatment of ischaemic stroke. The majority of genetic variants identified are associated with specific subtypes of ischaemic stroke, implying that these subtypes have distinct genetic architectures and pathophysiological mechanisms. The findings from the GWASs highlight the need to consider whether therapies should be subtype-specific. Further GWASs that include large cohorts are likely to provide further insights, and emerging technologies will complement and build on the GWAS findings.