A key feature of the Precision Nutrition and Health approach is the ability to tailor interventions to individual variability using multimodal data from large-scale biobanks and cohorts. Artificial intelligence (AI) and machine learning (ML) models offer new potential to model complex data but remain constrained by challenges related to data quality, interpretability, validation, and causal inference. This Perspective synthesizes current AI/ML methodologies in PN, elucidates their interplay with the distinctive features of multi-omic and nutritional data, such as being compositional, episodic, context-dependent, and error-prone, and delineates nutrition-specific best practices for achieving robust, interpretable, and clinically actionable AI integration in research and practice.
Bacteria possess an array of defenses against foreign invaders, including diverse nucleases that target and destroy the genome of an invading bacteriophage or foreign DNA element. A recently-described bacteriophage immunity pathway employs a cGAS/DncV-like nucleotidyltransferase to produce a cyclic tri-AMP second messenger, which activates the DNA endonuclease effector NucC. Here, we show that NucC is related to restriction enzymes but uniquely assembles into a homotrimer in solution. cAAA binding in a conserved allosteric pocket promotes assembly of two NucC trimers into a homohexamer competent for double-strand DNA cleavage. We propose that NucC mediates bacteriophage immunity either through global activation, causing altruistic cell death and abortive infection, or through local activation and targeted phage genome destruction. Finally, we identify NucC homologs in type III CRISPR-Cas systems, where they likely function as accessory nucleases activated by cyclic oligoadenylate second messengers synthesized by these systems’ effector complexes.
Microbially modified bile acids shape host physiology by regulating nutrient absorption, glucose homeostasis, circadian rhythms and thermoregulation. Here we identify a previously unrecognized drug-activating function of microbial bile acid conjugation. By systematically mining human LC–MS/MS datasets across public repositories and linking uncharacterized bile acid spectra to health-associated metadata, we discovered conjugates of the >75-year-old anti-inflammatory drug 5-aminosalicylic acid (5-ASA) with primary and secondary bile acids, including cholic, deoxycholic and lithocholic acids. These bile acid–drug conjugates were detected specifically in individuals treated with 5-ASA or its prodrugs. Multiple gut bacteria, including members of the Bacteroidota and Bacillota, generated cholyl–5-ASA in vitro, and bile salt hydrolase–associated transaminase activity was required for conjugate formation. In a mouse model of colitis, cholyl–5-ASA was associated with reduced intestinal inflammatory pathology and showed markedly enhanced activation of PPAR-γ in cell-based reporter assays compared with 5-ASA alone. Consistent with this activity, cholyl–5-ASA elicited selective immunophenotypic changes in CD4⁺ T cells in vitro, including increased Foxp3+ regulatory T cells. Together with prior evidence that 5-ASA efficacy depends on the microbiome, these findings support a model in which microbial bile acid conjugation represents a key activation step for 5-ASA therapy. More broadly, this work demonstrates how pan-repository metabolomics can uncover previously unrecognized microbiome-dependent chemical functions with direct therapeutic relevance.
1. Alzheimer’s disease (AD) is the most common neurodegenerative disease. AD is generally characterized by a progressive loss of neurons, extracellular accumulation of amyloid beta proteins, and in...