Observational epidemiology has implicated a wide range of biomarkers in the onset of human disease, however given that these methods are prone confounding it becomes difficult to identify causal biomarkers deserving of pharmaceutical intervention. Mendelian randomization (MR) is an emerging technique in human genetics which utilizes the power of random assignment of genetic information at meiosis to determine whether a biomarker is causal in human disease. It has been suggested that this technique can help guide drug development and inevitably improve the delivery of new drugs to clinical care. In this article we review examples of MR studies in cardiovascular disease and discuss limitations in the technique’s ability to predict therapeutic success in humans. (By Lauren E Mokry, http://jmg.bmj.com/content/early/2014/12/16/jmedgenet-2014-102438 )