Genomic and ecological systems-thinking framework for pathogenic Leptospira in Puerto Rico.

INTRODUCTION: Leptospirosis is a complex zoonotic disease requiring high-resolution surveillance. A systems-thinking framework was used to connect genomic and ecological data and map the geographic and host-based structuring of co-circulating pathogenic Leptospira lineages in Puerto Rico. METHODS: Forty-four core genomes of L. interrogans, L. borgpetersenii, and L. kirschneri from human, domestic, and wildlife hosts were analyzed. Spatiotemporal and landscape metadata were integrated using root-to-tip regression, isolation-by-distance profiling and calibrated single-nucleotide polymorphism (SNP) thresholds (≤1, ≤5, and ≤10 SNPs) to define transmission clusters. RESULTS: Leptospira species exhibited distinct ecological pathways partitioned by geography, explaining 56% of genomic variance for L. interrogans and 91% for L. borgpetersenii (PERMANOVA). L. interrogans displayed high landscape connectivity across multiple hosts, forming localized networks (≤1 to ≤10 SNPs) that capture active spillovers (human-to-rat linkages at ≤1 SNP) and resolved into rodent host-specific lineages (R2 = 0.34). Conversely, L. borgpetersenii showed spatial and temporal genomic homogeneity and a lack of host-associated structure within an unpartitioned transmission pool dominated by Mus musculus. As a result, fixed genomic thresholds yielded disparate outcomes: L. interrogans resolved into 4 to 5 discrete, expanding clusters, whereas L. borgpetersenii grouped into a single uniform population at the ≤10-SNP threshold. CONCLUSION: Co-circulating pathogenic leptospires occupy distinct ecological niches shaped by varying host restriction and environmental persistence. Fixed genomic thresholds lack universal applicability; effective genomic epidemiological surveillance must employ species-specific threshold calibration to accurately map transmission pathways.