BACKGROUND: Bacterial vaginosis (BV) is an enigmatic disease of unknown origin that affects a large percentage of women. The vaginal microbiota of women with BV is associated with serious sequelae, including abnormal pregnancies. The etiology of BV is not fully understood, however, it has been suggested that it is transmissible, and that G. vaginalis may be an etiological agent. Studies using enzymatic assays to define G. vaginalis biotypes, as well as more recent genomic comparisons of G. vaginalis isolates from symptomatic and asymptomatic women, suggest that particular G. vaginalis strains may play a key role in the pathogenesis of BV. METHODOLOGY/PRINCIPAL FINDINGS: To explore G. vaginalis diversity, distribution and sexual transmission, we developed a Shannon entropy-based method to analyze low-level sequence variation in 65,710 G. vaginalis 16S rRNA gene segments that were PCR-amplified from vaginal samples of 53 monogamous women and from urethral and penile skin samples of their male partners. We observed a high degree of low-level diversity among G. vaginalis sequences with a total of 46 unique sequence variants (oligotypes), and also found strong correlations of these oligotypes between sexual partners. Even though Gram stain-defined normal and some Gram stain-defined intermediate oligotype profiles clustered together in UniFrac analysis, no single G. vaginalis oligotype was found to be specific to BV or normal vaginal samples. CONCLUSIONS: This study describes a novel method for investigating G. vaginalis diversity at a low level of taxonomic discrimination. The findings support cultivation-based studies that indicate sexual partners harbor the same strains of G. vaginalis. This study also highlights the fact that a few, reproducible nucleotide variations within the 16S rRNA gene can reveal clinical or epidemiological associations that would be missed by genus-level or species-level categorization of 16S rRNA data.