Editors' ChoiceNutrition

Personalized dietary advice on the horizon

See allHide authors and affiliations

Science Translational Medicine  24 Jun 2020:
Vol. 12, Issue 549, eabc8949
DOI: 10.1126/scitranslmed.abc8949

Abstract

Multimodal measurements and machine learning provide evidence for personalized dietary advice.

A substantial proportion of global mortality and morbidity due to lifestyle-related diseases is attributable to poor nutrition. Although interventions to change food supply chains on a public health level are likely to have the most impact, dietary advice tailored to an individual’s physiology and environment also holds great promise. This type of personalized dietary advice is likely to benefit from digital and mobile solutions. However, rigorous evidence quantifying individual variations in response to different meal compositions has been mostly lacking until now.

Berry et al. demonstrated that there are large differences in how individuals respond to different meals using a combination of in-clinic, at-home, and mobile measurements in 1000 individuals from a well-characterized cohort in the United Kingdom. Briefly, they studied metabolic responses to meals of differing macronutrient composition and correlated the meal compositions with biochemical measurements, continuous glucose monitoring, genetics, microbiome, and lifestyle characteristics. The variation in interindividual responses could be explained more by environmental (meal composition and time of meal) as opposed to genetic (microbiome and genome) factors. Going a step further, they created a machine learning algorithm to leverage the collected data to predict dietary responses in an independent group of 100 individuals, showing good performance.

This study shows the importance of incorporating multiple data types for accurate prediction of individual dietary responses and may form the basis for personalized dietary guidance in the future. In addition, this study also highlights the limited role that genomic factors may play in dietary response. This work is particularly relevant in light of previous data showing that individualized nutritional advice may be preferable over generalized nutritional recommendations. Although it is too early to recommend population-wide personalized guidance for nutrition, and randomized controlled trials are still required, this work provides rigorous evidence supporting personalized nutrition at a population level.

Highlighted Article

View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article