Salivary Dipeptides and Microbiota Multiomics Biomarkers of Caries

Artikel i vetenskaplig tidskrift, 2026

Dental caries is a prevalent disease shaped by dietary and metabolic interactions between the host and oral microbiota. While separate microbiome and metabolome profiling have improved the understanding of caries etiology, integrated multiomics approaches remain underutilized. We aimed to identify salivary biomarkers of caries experience in a cross-sectional study combining microbiome, metabolome, sugarome, and diet, oral, and demographic information in 217 healthy Swedish adolescents and young adults. Participants underwent clinical examination, dietary assessment, and saliva analysis using full-length 16S rRNA gene sequencing, untargeted liquid chromatography-mass spectrometry (LC-MS) metabolomics, and 2-dimensional gas chromatography-MS (GC & times;GC-MS)-based sugar profiling. Caries signs and restorations were summarized as decayed and filled surfaces including enamel caries. Associations between caries experience, demographics, and omics measures were evaluated using a nested cross-validated linear-regression framework with inner-loop feature selection and outer-loop performance assessment, emphasizing effect-direction stability and reproducibility across cross-validation models. Selected OMICs features were then integrated using orthogonal partial least-squares (OPLS) modeling. Three salivary dipeptides (Ser-Pro, Phe-His, and Thr-Gly) showed robust inverse associations with caries experience. Several cariogenic taxa, including Streptococcus mutans, Streptococcus sobrinus, Lactobacillus gasseri, and selected Actinomyces and Prevotella species, were positively associated with caries experience, whereas Veillonella rogosae, Stomatobaculum sp. HMT-097, and other taxa as well as 5 bacteria-predicted enzymes were associated with lower caries experience. Dietary variables and salivary sugars showed no consistent associations with caries experience. An integrated OPLS model incorporating demographics, microbiota composition, microbial-predicted enzymes, and metabolomics explained 41% of the variation in caries (R 2 = 0.41) with a cross-validated predictive ability of 30% (Q 2 = 0.30). This saliva-based multiomics framework revealed coherent microbial and metabolic signatures of caries experience. The Ser-Pro dipeptide showed the strongest inverse association and warrants validation in longitudinal cohorts to establish temporal precedence and clinical utility as a novel caries biomarker.