Background The cervicovaginal mucosa is inhabited by an ecosystem of bacteria, fungi and viruses, which likely interact with each other. The vaginal virome may influence vaginal immunity directly, or though modulation of the bacterial component via bacterial-phage dynamics. These interactions may play an important role in sexual and reproductive health outcomes. However, it is predicted that over 60% of the human DNA virome has not yet been identified, and the RNA virome is even less explored, nor has the impact of HIV on the vaginal virome composition been described.

Methods We optimised viral particle extraction from vaginal swabs. We then optimised a Sequence Independent Single Primer Amplification (SISPA) approach to enable deep sequencing of the viral metagenome that decreases the GC and genome size bias introduced by commonly used methods such as Multiple Displacement Amplification and Rolling Circle Amplification, while also yielding a greater diversity of near-complete metagenome-assembled genomes.

Results SISPA was able to recapitulate almost exactly the relative abundance of a viral mock community. Storing swabs in universal transport media (UTM) directly after collection and treating the sample with a 18G needle prior to viral particle extraction resulted in the greatest yield of viral nucleic acid and subsequent read depth, over storing swabs dry or in a 1:1 dilution of UTM and SM buffer. We have applied this method to vaginal swabs from Sub-Saharan African women with and without HIV.

Conclusion This work lays the ground work for project TMA2020CDF-3192, which will assess the interaction of vaginal virome and bacteriome in pregnant women with HIV in sub-Saharan Africa and risk of preterm birth.