About 70 years ago, Ancel Keys proposed the diet and heart hypothesis. However, despite its overall success, this hypothesis has been continuously disputed. Why something so simple as the diet-CVD relationship, got so complicated?
Because we are complicated, as shown by the huge interindividual variability and low intraindividual reproducibility observed in intervention studies. Moreover, food is complicated, and there are still thousands of functional compounds, unknown to us, that could be key to understand how what we eat affects our health.  
In the '80s, the ability to detect genetic variants gave birth to another simple hypothesis: variability in genes drives interindividual variability in diet response.  The testing of this hypothesis applied to CVD has resulted in over 1000 publications. However, most have investigated gene-diet interactions related to CVD risk factors, and very few have targeted CVD events. The Predimed study is among the few intervention studies describing gene-diet interactions for CVD events, as shown by research involving the LPL and TCF7L2 genes. 
Another source of complexity is time. It is important what we eat, but also when we eat. This may be highly relevant for CVD that has been shown to have circadian and seasonal rhythmicity. In this regard, we have demonstrated that genetic variants at the circadian CLOCK gene modulate the relation between diet and CVD events.
After decades focusing on gene-diet interactions, precision nutrition has grown to incorporate other players, such as epigenetics, metabolomics, microbiome, aging, sex, lifestyle, chronobiology, diseases, drugs, and environment. Given the vast increase in data collected, we must adopt new approaches to query and analyze the Big Data sets, such as Artificial Intelligence. This combination of data and analytical tools will pave the way to precision nutrition and successful CVD prevention and therapy.