Abstract
Limosilactobacillus fermentum is a widely distributed lactic acid bacterium frequently detected in fermented foods and host-associated microbiota, yet its global genomic diversity and functional variability remain insufficiently characterized. Here, we performed a large-scale comparative genomic analysis of 336 high-quality L. fermentum genomes curated from public databases. Species identity was validated using average nucleotide identity (ANI), and population structure was examined using pairwise ANI comparisons together with Mash-based phylogenetic reconstruction. Clustering at ≥ 99% ANI resolved the dataset into 15 genomic clusters, with four dominant lineages comprising the majority of genomes. Pangenome reconstruction identified 5,853 gene clusters, including 1,325 core genes (22.6%) and a large accessory component dominated by low-frequency genes. Heap's law modeling (λ = 0.19) indicated a weakly open pangenome, suggesting ongoing gene acquisition as additional genomes are sampled. Functional annotation revealed that core genes were primarily associated with essential cellular processes, whereas accessory genes were enriched in carbohydrate metabolism, membrane-associated functions, and defense-related systems. Variation in carbohydrate-active enzymes (CAZymes), transport systems, and stress-response genes was observed across lineages, indicating strain-level functional diversity. Although genomes from human and food sources were broadly distributed across phylogenetic lineages, multivariate analysis showed that gene-content variation was more strongly associated with genomic lineage than with isolation source. These results provide a population genomic framework for understanding genomic diversity and functional potential in L. fermentum.