The global rise in childhood obesity and metabolic disorders such as type 2 diabetes has closely paralleled industrialization and increased chemical emissions. While genetics, diet, and physical activity are important contributing factors, they do not fully explain the rapid surge in these conditions. Experimental evidence suggests that exposure to environmental chemicals—particularly during fetal development and early childhood—can disrupt metabolism and increase the risk of weight gain, with lasting effects. However, the underlying mechanisms in humans are still not well understood, and the potential role of the gut microbiome in this process remains largely unexplored. The NEMESIS project aims to investigate how endocrine-disrupting chemicals (EDCs) contribute to metabolic dysfunction. Our subproject specifically focuses on the gut microbiome’s role in mediating these effects. We aim to identify how EDC exposure alters gut microbial composition and gene function by analyzing human cohort samples and conducting exposure studies in mice and zebrafish. To achieve this, we will perform metagenomic sequencing (shotgun or long-read) on total DNA extracted from intestinal or fecal samples. Standardized univariate and multivariate statistical methods will be used to analyze microbial profiles. Feature selection and classification approaches will help us identify microbial biomarkers that predict EDC-induced metabolic disturbances. These microbial data will then be integrated with metabolomics data to uncover significant interactions between microbial taxa and host metabolites. To explore these relationships in depth, we will apply unsupervised machine learning techniques—such as regularized canonical correlation analysis and joint non-negative matrix factorization—to detect patterns linking metabolomic and microbial profiles. We will refine our biomarker selection to focus on those specifically associated with key metabolic outcomes. Finally, to directly assess the role of the microbiome in mediating EDC-induced metabolic effects, we will generate germ-free zebrafish models by incubating embryos (2–6 days post-fertilization) in sterile, antibiotic-supplemented medium, followed by toxicity screening.