Classifying Individuals Into a Dietary Pattern Based on Metabolomic Data.

Abstract

The objectives are to develop a metabolomic-based model capable of classifying individuals into dietary patterns and to investigate the reproducibility of the model. K-means cluster analysis is employed to derive dietary patterns using metabolomic data. Differences across the dietary patterns are examined using nutrient biomarkers. The model is used to assign individuals to a dietary pattern in an independent cohort, A-DIET Confirm (n = 175) at four time points. The stability of participants to a dietary pattern is assessed. Four dietary patterns are derived: moderately unhealthy, convenience, moderately healthy, and prudent. The moderately unhealthy and convenience patterns has lower adherence to the alternative healthy eating index (AHEI) and the alternative mediterranean diet score (AMDS) compared to the moderately healthy and prudent patterns (AHEI = 24.5 and 22.9 vs 26.7 and 28.4, p <0.001). The dietary patterns are replicated in A-DIET Confirm, with good reproducibility across four time points. The stability of participants' dietary pattern membership ranged from 25.0% to 61.5%. The multivariate model classifies individuals into dietary patterns based on metabolomic data. In an independent cohort, the model classifies individuals into dietary patterns at multiple time points furthering the potential of such an approach for nutrition research.