The findings establish an association within this small, geographically localized cohort, but they do not show that oral bacteria cause obesity or that the observed taxa can predict it in other children. The study, published in International Microbiology, was designed to examine both the structure and functional potential of the oral microbiome. Its authors argue that previous work on childhood obesity and oral microbes has been limited by reliance on 16S ribosomal RNA sequencing, a method that can provide broad taxonomic profiles but has limited species-level and functional resolution.

This is not a direct measurement of every organism’s activity. The genes detected indicate biological potential, and their expression or effect in a child was not measured. Participants were recruited in Gansu Province, and a doctor measured each child’s height and weight.

Children with obesity were identified using Chinese criteria, while those who had taken antibiotics within the preceding three months, had diseases affecting weight, or had drug-induced obesity were excluded. The two groups did not differ significantly in sex, age or height, although body weight and body mass index differed as expected from the group definitions. Before collection, participants abstained from food for two hours.

The researchers then collected 4 millilitres of unstimulated saliva and stored it at minus 80 degrees Celsius until DNA extraction. Sequencing was performed on an Illumina NovaSeq PE150 platform. After adapters and low-quality reads were removed and human-derived sequences were filtered out, the investigators retained 798,884,160 high-quality reads from 804,925,436 raw reads, with an average of 44,382,453 reads per sample, ranging from 43,182,540 to 45,964,730.

They built a gene catalogue by clustering predicted gene sequences at 90% sequence identity and 90% coverage, then used the DIAMOND alignment tool to compare those sequences with the NR database for taxonomic annotation and with the Kyoto Encyclopedia of Genes and Genomes for functional annotation. Relative abundances were calculated from the resulting alignments. The study analyzed 125 phyla, 185 classes, 300 orders, 572 families, 1,852 genera and 7,459 species across the 18 samples.

At the broadest level, the oral communities of the two groups were similar. Bacteroidetes, Firmicutes, Proteobacteria, Fusobacteria and Actinobacteria together accounted for 96.2% of the microbiota in normal-weight children and 94.5% in children with obesity. The ratio of Firmicutes to Bacteroidetes was numerically higher in the obese group, at 0.97 compared with 0.77, but that difference did not reach statistical significance, with P greater than 0.05 in the Wilcoxon rank-sum test.

Shannon and Simpson indices, which summarize within-sample diversity, likewise showed no significant group difference. Principal coordinate analysis and non-metric multidimensional scaling based on species-level Bray–Curtis distances also found no significant separation between the groups. These results argue against describing childhood obesity here as a wholesale transformation of oral microbial diversity.

More detailed taxonomic comparisons nevertheless identified 12 species with significantly different relative abundances. Six were enriched in normal-weight children: Dialister invisus, Treponema porcinum, Moraxella sp., Moraxella sp. VT 16 12, Moraxella lacunata and Moraxella caprae.

The other six were enriched in children with obesity: Campylobacter concisus, Veillonella sp. oral taxon 158, Veillonella rogosae, Veillonella sp. S13053 19, Streptococcus ilei and Streptococcus australis. At the genus level, Moraxella and Dialister were more abundant in the normal-weight group.

The researchers used linear discriminant analysis effect size to screen for taxa that might distinguish the groups, retaining a threshold above 2, and assessed abundance differences with Wilcoxon rank-sum tests at a significance level below 0.05. A species-level Venn analysis showed that 6,223 of the 7,459 observed species were shared between the groups, representing 83.4% of all species detected. The researchers counted 725 species unique to the obese group and 511 unique to the normal-weight group.

The pattern is consistent with substantial overlap in the oral microbiota, with differences concentrated in the relative abundance of particular taxa rather than in a large separation of the overall communities. The authors note that oral microbial composition can vary with ethnicity, diet and geographical region, which is one reason they treated the findings as population-specific observations requiring further study rather than universal markers. Metagenomic functional annotation produced 231,437 genes assigned to six primary KEGG pathway categories.

The largest was the general metabolic pathway, representing 44.5% of annotated genes and 120,029 genes, followed at the secondary level by translation at 11.9% and 12,736 genes. Other relatively abundant categories included carbohydrate metabolism at 8.9%, cofactor and vitamin metabolism at 8.0%, amino acid metabolism at 7.6%, energy metabolism at 6.3%, membrane transport at 6.0%, nucleotide metabolism at 5.5% and replication and repair at 5.0%. At the more detailed KEGG level, 16 pathways differed significantly between the groups, with a significance threshold of P below 0.05.

The normal-weight group had greater relative abundances of genes assigned to hypoxia-inducible factor-1 signalling, peroxisome proliferator-activated receptor signalling, beta-alanine metabolism, AMP-activated protein kinase signalling and apoptosis. Children with obesity instead showed greater relative abundances of genes associated with biofilm formation–Vibrio cholerae, selenocompound metabolism, RNA polymerase, xylene degradation and dioxin degradation. Four obesity-enriched species—Veillonella sp. oral taxon 158, Veillonella sp.

S13053 19, Streptococcus ilei and Streptococcus australis—were strongly positively correlated with the biofilm, selenocompound, xylene-degradation and dioxin-degradation pathways. These are correlations within the dataset, not evidence that the bacteria caused the pathways to change or that the pathways mediate obesity. The pathway labels also describe inferred functions rather than confirmed processes, and a gene associated with a named species in the KEGG database does not demonstrate that the child’s oral community carried out that activity.

The findings are biologically plausible in some respects but remain preliminary. Veillonella and Streptococcus are common early colonizers of dental plaque and can interact while biofilms develop. The study’s authors also note that C. concisus produces zonula occludens toxin, which has been associated with effects on intestinal epithelial permeability, and that oral bacteria and their products may reach distant tissues through swallowing, airflow, blood circulation and neural pathways.

These observations provide possible routes for investigation, but the present experiment did not measure intestinal permeability, systemic inflammation, swallowed organisms or the transfer of oral microbes to the gut. The detected differences in selenocompound metabolism and metabolic-signalling genes may be biologically relevant, yet the authors do not directly demonstrate a pathway connecting oral microbial function to childhood obesity. Several limitations constrain how far the results can be generalized.

The sample included only 18 children, all recruited within one province, and the groups had an imbalanced sex ratio. Dental caries, gingivitis, diet and oral-hygiene habits were not systematically recorded, so possible confounders remain. The cross-sectional design also prevents the researchers from determining whether microbial differences preceded weight gain, developed after obesity was established, or merely accompanied it.

The authors call for larger, longitudinal studies with comprehensive clinical and oral-health measurements. Taken together, the study provides a species-level and functional map of oral microbial differences in a small pediatric cohort. It identifies candidate taxa and pathways for future testing, but it neither establishes a microbial cause of childhood obesity nor supports clinical use of oral metagenomic signatures.

Subject of Research: Oral microbiota composition and functional pathways in children with obesity

Article Title: Metagenomic analysis of oral microbiota in Chinese children with obesity

Article References: Wang, W., Feng, Y., Wang, Y., Wang, G., & Wang, X. (2026). Metagenomic analysis of oral microbiota in Chinese children with obesity. International Microbiology. https://doi.org/10.1007/s10123-026-00913-1

Image Credits: AI Generated

DOI: 10.1007/s10123-026-00913-1

Keywords: oral microbiota, childhood obesity, metagenomics, saliva, pediatric health, microbial diversity, biofilm formation, KEGG pathways, China, Metagenomic, analysis, oral