A peer-reviewed survey assembled more than a million viral groups from thousands of livestock-manure datasets. Established antibiotic-resistance gene types were uncommon on viral sequences, although predicted links connected viruses with resistant bacteria. The catalogue expands a poorly mapped microbial environment; the computational associations do not demonstrate that viruses transferred resistance genes.
A single peer-reviewed genomic survey assembled 1,330,372 viral groups from 4,017 public metagenomes of chicken, swine and cattle manure. These groups approximate distinct viral lineages. More than 99% were absent from the comparison databases, expanding the reference material available for a microbial environment where antibiotic resistance is a concern.[1]
The viral sequences contained 78 subtypes of established antibiotic-resistance genes, compared with 857 in the whole sampled resistance-gene pool. The resulting 9.1% describes the share of gene types represented, rather than the proportion of viruses carrying resistance. The wider resistance pool included genes such as mcr, blaNDM and blaCTX-M, which raise clinical concerns. None of these appeared in the viral sequences examined.[1]
Only 831 viral groups occurred in all three manure types. Testing two independent manure studies also showed that the specialised catalogue captured more viral sequence reads than several general databases. Such comparisons measure the catalogue’s usefulness for identifying sequences, without observing a virus infecting a host.[1]
Sequence comparisons and other computational signals linked viruses to resistant bacteria. Around two-thirds of resistance-bearing viral fragments also carried mobile genetic elements, suggesting possible movement. Assembly, gene annotation and host prediction each retain uncertainty. Laboratory validation is needed to establish active transfer, and the survey did not directly measure bacterial DNA packaged separately from a viral genome.[1]
