The Framingham Heart Study (FHS) was established in 1948 to improve understanding of the epidemiology of coronary heart disease (CHD) in the USA. In 1961, seminal work identified major risk factors for CHD (high blood pressure, high cholesterol levels and evidence on the electrocardiogram of left ventricular hypertrophy), which later formed the basis for multivariable 10-year and 30-year risk-prediction algorithms. The FHS cohorts now comprise three generations of participants (n ≈ 15,000) and two minority cohorts. The FHS cohorts are densely phenotyped, with recurring follow-up examinations and surveillance for cardiovascular and non-cardiovascular end points. Assessment of subclinical disease and physiological profiling of these cohorts (with the use of echocardiography, ambulatory electrocardiographic monitoring, exercise stress testing, cardiac CT, heart and brain MRI, serial vascular tonometry and accelerometry) have been performed repeatedly. Over the past decade, the FHS cohorts have undergone deep 'omics' profiling (including whole-genome sequencing, DNA methylation analysis, transcriptomics, high-throughput proteomics and metabolomics, and microbiome studies). The FHS is a rich, longitudinal, transgenerational and deeply phenotyped cohort study with a sustained focus on state-of-the-art epidemiological methods and technological advances to facilitate scientific discoveries.