Taxonomic identification and characterization of Leptospira from retrospective clinical samples in the Philippines

Leptospirosis is a highly endemic neglected tropical disease in the Philippines, where conventional diagnostic methods such as the microscopic agglutination test (MAT) provide limited information on circulating Leptospira and are affected by serological cross-reactivity. This study evaluated the utility of shotgun metagenomic next-generation sequencing (mNGS) for pathogen detection and species-level characterization using 20 archived serum-derived total nucleic acid extracts collected from suspected leptospirosis cases. Samples were previously tested by quantitative real-time PCR (qPCR) and MAT, and sequenced using the Illumina MiSeq platform. Host-derived reads were removed using BWA and SAMtools prior to taxonomic profiling with KrakenUniq. Species-level Leptospira assignments were retained based on predefined KrakenUniq filtering criteria, including a minimum taxReads threshold and evidence of taxon-specific unique k-mers. Leptospira taxReads were detected in 58.8% (n=10/17) of qPCR-positive samples. Notably, mNGS detected pathogenic Leptospira in 60% (n=6/10) qPCR-positive but MAT-negative samples, demonstrating its potential to complement conventional serology for etiological characterization. MAT reactivity was predominantly against L. biflexa serovar Patoc (n=4/5), with one sample reacting to L. interrogans serovar Autumnalis in the MAT-positive samples. Meanwhile, mNGS identified a predominance of L. interrogans, with additional detection of L. borgpetersenii and L. kirschneri. The absence of detectable Leptospira in some qPCR-positive samples underscores the sensitivity limitations of mNGS in low-burden or archived samples. Overall, mNGS provided species-level information beyond conventional diagnostics and supports its potential application in Leptospira genomic surveillance in the Philippines.