Personalized Nutrition by Prediction of Glycemic Responses.

A machine-learning algorithm is devised that integrates blood parameters, dietary habits, anthropometrics, physical activity, and gut microbiota measured in an 800-person cohort and shows that it accurately predicts personalized postprandial glycemic response to real-life meals, and a blinded randomized controlled dietary intervention based on this algorithm resulted in significantly lower postpr andial responses and consistent alterations to gut microbiota configuration.