Connections of microbiome composition with host metabolism and habitual diet: the results of PREDICT 1 study
In this manuscript Asnicar and colleagues described the results of Personalised Responses to Dietary Composition Trial (PREDICT 1), a study that included 1098 individuals profiled pre and post standardized dietary challenges using intensive in-clinic biometric and blood measures, habitual dietary data collection, continuous glucose monitoring and stool metagenomics. The main objective of this trial was to quantify and predict individual variations in metabolic responses to standardized meals: data from a cohort of twins and unrelated adults were integrated in order to explore genetic, metabolic, microbiome composition, meal composition and meal context data to distinguish predictors of individual responses to meals.
First, the authors leveraged a subpopulation of mono- and dizygotic twins and confirmed that host genetics had a limited influence on the composition of the microbiome. Then, they investigated the intrasample microbiome diversity (alpha diversity) and observed that was positively or negatively associated with several cardiometabolic biomarkers, like high-density lipoprotein cholesterol (HDLC) and glycoprotein acetyl A (GlycA), and diet. These results prompted the authors to assess the links between habitual diet and microbiome using random forest models, each trained on quantitative microbiome features to predict each dietary variable from food frequency questionnaires. Constituent foods were summarized into dietary indices like Healthy Food Diversity, Healthy/Unhealthy Plant-based Dietary Indices, Healthy Eating Index and alternate Mediterranean Diet: a strong correlation was demonstrated between these indices and microbial composition, highlighting the relationship between microbiome and health associated dietary patterns. Microbial taxa most responsible for these diet-based community associations segregate in two clusters: taxa linked to healthy plant-based foods (butyrate producers) and taxa linked to several less healthy plant-based and animal-based foods (Clostridium species). However many of the strongest microbial associations with diet occurred with only recently isolated or still uncultured taxa. Then, using a machine learning approach, the authors found that visceral fat and cardiometabolic markers were strongly linked with microbial composition. They also observed a consistent set of microbial species that were strongly linked to postprandial responses and to fasting circulating metabolites connected with cardiometabolic risk.
The importance of the PREDICT 1 study is that is the first one to identify a shared diet-metabolic health microbial signature, segregating favorable and unfavorable taxa with various measures of both dietary intake and cardiometabolic health. The results of this study will be useful in the utilization of gut microbiome as a biomarker for cardiometabolic risk and in strategies for modify microbiome composition in order to improve personalized dietary health.