Q 2 cum, model predictive ability parameter according to cross validation; ROC, receiver operating curve; R 2 X, goodness-of-fit parameter; R 2 Y, proportion of the variance of the response variable that is explained by the model; T0, baseline time; T3, after 12-w of nuts consumption; VIP, variable importance for projection; XIC, extracted ion chromatogram.Keywords: adiposity / biomarker of nuts / gut microbiota / metabolomics / plasma human ABSTRACT Scope. To identify the most discriminant dietary biomarkers of nuts exposure in subjects with metabolic syndrome (MetS), and investigate the potential association between exposure and the severity of the MetS diagnostic traits. Methods and results. We applied the untargeted LC-ESI-qToF-MS-driven metabolomic workflow to explore the changes occurring in the plasma metabolome of MetS subjects following 12-week intake of mixed nuts (30 g/d) (nuts versus control groups). Urolithin A glucuronide was the most discriminative biomarker of nut exposure, showing the highest predictive capacity [ROC AUC = 89.6% (80.8-98.4)] despite the inter-individual variation expected for a host-microbial cometabolite. Furthermore, the detection of urolithin A glucuronide in plasma showed significant inverse correlation with basal abdominal adiposity (waist circumference: r=-0.550, p<0.01; waist-hip ratio: r=-0.409, p<0.05) and impaired glycemic control (fasting insulin: r=-0.414 & HOMA-IR: r = -0.417, p<0.05). Significant changes in mediumchain dicarboxylic acids, recognized as alternative energy substrates that are particularly 2 relevant in the case of glycemic control impairment, were also associated with nut consumption. Conclusions. Higher levels of urolithin are reported in subjects with less severe MetS traits, especially in females. We believe that this inverse correlation may be related with profile of gut microbial dysbiosis, recently associated to subjects with MetS.