Characterizing the parvovirome of Swazi bats: novel species, highly divergent lineages, endogenous viral elements, and taxonomic challenges.

Parvoviridae (small, nonenveloped ssDNA viruses) currently includes 281 species in two vertebrate- and four invertebrate-infecting subfamilies. While parvovirus-derived sequences are frequently identified in viromes, their taxonomy and host affiliation can be challenging due to high host and genetic diversity. We investigated the faecal parvovirome of 46 bats (7 insectivorous and 1 frugivorous species) from Eswatini and identified 28 novel viral species in 29 individuals (63.0%). The majority of these (22/28, 78.6%) belonged to nine genera (including two that are previously undescribed) within the invertebrate-infecting subfamily Densovirinae. A novel virus in the genus Brevipenbrevirus (arthropod-infecting subfamily Penbrevirinae) was found in 19.6% of the animals, including several frugivorous Epomophorus wahlbergi bats. A novel bat protoparvovirus (vertebrate-infecting subfamily Parvovirnae) was found both in the faeces and blood of one Afronycteris nanus bat. A highly divergent virus (Swazi bat-associated megaparvovirus 1, SwaBA-MePV-1) was found in the faeces, but not in the blood, of two insectivorous bats (Mops pumilus and Scotophilus viridis). Compared to other parvoviruses, SwaBA-MePV-1 presented two additional coding cassettes, significantly increasing its genome size. Homology modelling showed capsid protein C-terminal elongation, a previously undescribed strategy of parvoviral particle size expansion. Exploring public repositories identified 10 related uncharacterized viruses with similar genome organization and complete endogenous viral elements (EVEs) in eight beetle species, suggesting a coleopteran host affiliation. The complete genome of another highly divergent virus (SwaBA microparvovirus 1), without any detectable exogenous or endogenous relatives, was found in the faeces, but not in the blood, of one insectivorous Mops condylurus bat. Importantly, when comparing our sequences to references in Genbank, we observed that taxonomic mislabelling in sequence repositories can seriously misguide automatic taxonomy assignments (~75% of sequences initially identified as parvoviruses were discarded as false positives). These errors are amplified as new mislabelled sequences become dominant, highlighting the importance of prioritizing taxonomy validation and correct annotations in repositories. This study demonstrates that faecal samples from insectivorous chiropterans are rich in (novel) parvoviruses from various hosts. The discovery of highly divergent lineages (outside current sub-families) and EVEs helps clarify parvovirus evolutionary history and emphasizes how much of the parvoviral world remains unexplored.