Table 2.

Alterations in pathogen-specific miRNA expression and their functional implications in periodontitis.

Pathogen	Model system	Time points	Key miRNA findings	Machine learning identified	Limitations	REF.
S. gordonii	C57BL/6J mice	8 and 16 weeks	At 8 weeks: 191 miRNAs upregulated, 22 downregulated; at 16 weeks: 10 upregulated, 32 downregulated; miR-210 and miR-423-5p consistently downregulated	miR-22, miR-30c (8wk); miR-720, miR-339-5p (16wk)	Single-species model; does not reflect human polymicrobial biofilm	(37)
P. gingivalis	C57BL/6J mice	8 and 16 weeks	At 8 weeks: 26 upregulated, 14 downregulated; at 16 weeks: 7 upregulated, 1 downregulated; miR-103 and miR-30d upregulated at both time points	Not reported	In vitro LPS-induced miRNAs not detected in vivo; human validation limited	(51)
P. gingivalis + T2DM	Human gingival tissue (64 patients)	Cross-sectional	miR-155 upregulated 11.17-fold in periodontitis+T2DM; M1 polarization	AUC = 0.74 (TNF-α)	Cross-sectional; modest diagnostic efficacy for miR-155	(52)
T. forsythia	C57BL/6J mice	8 and 16 weeks	At 8 weeks: 67 downregulated (miR-375, miR-200c, miR-200b); at 16 weeks: 16 upregulated (miR-let-7c, miR-146a), 32 downregulated	miR-592-5p, miR-339-5p	Single-species model; machine learning findings lack external validation	(54)