Machine learning-integrated full-length 16S rRNA gene profiling reveals early-life gut microbial signatures associated with subsequent infantile colic in Thai infants
Year: 2026
Type: article
Study design: Observational Study
Abstract: Infantile colic develops after the neonatal period, but neonatal gut microbial patterns preceding symptom onset remain unclear. We examined early-life gut microbiome associations with subsequent infantile colic and internally evaluated microbiome-only prediction alongside retrospective metadata-containing comparisons. Stool samples collected 0-4 days after birth from 209 Thai infants (119 healthy and 90 who later met Rome IV diagnostic criteria for infantile colic) were profiled by Oxford Nanopore full-length 16S ribosomal RNA (rRNA) gene sequencing. Analyses included diversity and composition assessments, covariate-adjusted MaAsLin3 models, repeated nested cross-validation, and exploratory conditional association networks. Infants who subsequently developed infantile colic had lower Chao1 richness than healthy infants (q = 8.7 × 10−5), although delivery mode and collection day explained more community variation than outcome status. After covariate adjustment, Enterococcus faecium presence was associated with subsequent colic (odds ratio, 4.23; q = 0.021). The fully nested model-development procedure yielded participant-level area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) values of 0.773 and 0.740. In a secondary fixed-model comparison including feeding summarized through one month, a species-presence radial basis function support vector machine outperformed metadata alone (mean outer-fold AUROC/AUPRC, 0.780/0.764 vs. 0.539/0.484), while adding metadata produced only a small, uncertain improvement. Exploratory network analysis contextualized candidate taxa within group-specific conditional association modules. These findings support a focused early-life species-level pattern associated with subsequent infantile colic and prioritize candidate microbial features for external validation.
Source: BioScience Trends
Institution: Chulalongkorn University
Language: English
Topic: Gut microbiota and health
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: diamond
OpenAlex ID: W7235043806
