<h3>Background</h3> Large-scale genomic analyses of Treponema pallidum (Tp) are now possible using enrichment methods that enable sequencing without in vivo cultivation in rabbits. Using capture and amplification enrichment approaches, we are generating complete genomes from diverse clinical samples collected as part of a multi-site syphilis vaccine development project. Analysis of these and a growing number of publicly available genomes provides new insights into the global diversity and evolution of Tp. <h3>Methods</h3> Clinical samples were collected from subjects presenting with primary and secondary syphilis in the US, Columbia, Malawi, China, and Czech Republic. DNA extracted from ulcer swabs, skin biopsies, and rabbit-passaged clinical isolates was enriched for Tp using pooled, parallel whole-genome amplification and/or RNA baits before Illumina short-read sequencing. We adopted hybrid comparative genome assembly analysis that involved de novo and reference-based assembly strategies to generate whole-genome consensus sequences. Repetitive and paralogous loci are being finished using Oxford Nanopore Technologies long-read sequencing and outer membrane protein (OMP) variability assessed using three-dimensional structural modelling. <h3>Results</h3> We whole-genome sequenced 20 newly collected strains and analyzed publicly available sequencing data from 605 Tp samples. We observed a higher proportion of Nichols-like strains (6 of 20 [30%]) than expected in newly sequenced samples. Macrolide resistance-associated 23S rRNA alleles were not observed in samples from Columbia or Malawi. Initial analysis of bamA, lptD, and fadL-like loci suggests a spectrum of OMP variability that supersedes clade assignment. Analyses of newly collected samples and publicly available genomes is ongoing. <h3>Conclusion</h3> Sampling from geographically diverse sites suggests that Tp strains are more genetically diverse than previously appreciated. Nichols and SS-14 clades were well represented, and macrolide susceptibility was maintained in some localities. Representative sampling complemented by analysis of robust clinical data and paired biological samples is critical to the development of a globally effective syphilis vaccine.