AI-Powered Discovery of Novel RNA Viruses from the Permafrost of a 14,300-Year-Old Pleistocene Wolf

Ancient viruses preserved as molecular relics offer rare and often unpredictable insights into virus-host co-evolution and the ecological dynamics of past ecosystems. However, the recovery of ancient RNA viruses via paleotranscriptomics has remained largely unexplored, constrained by the inherent chemical instability of RNA and the lack of sensitive detection tools capable of identifying deeply divergent sequences. Here, we leveraged recent advances in artificial intelligence and high-throughput sequencing to conduct comprehensive metatranscriptomic mining of publicly available RNA-seq datasets from three ancient or extinct host species: the woolly mammoth (Mammuthus primigenius), the Tasmanian tiger (Thylacinus cynocephalus), and the gray wolf (Canis lupus). We performed sensitive homology searches using the AI-driven protein language model Lucaprot, coupled with structural validation via AlphaFold2, to screen billions of raw sequencing reads for conserved viral RNA-dependent RNA polymerase (RdRp) signature genes. Our pipeline identified two near-complete previously unknown RNA viruses in a 14,300-year-old Pleistocene wolf specimen. Phylogenetic analyses placed them within established mycovirus genera (Duamitovirus and Orthocurvulavirus), indicating they infected fungi that inhabited the carcass rather than the wolf itself. Despite deep sequence divergence from known viruses (57.6% and 60.8% RdRp amino acid identity, respectively), the catalytic A, B, and C motifs remain structurally intact. Strict authentication through multiple analyses firmly verified their ancient provenance. To our knowledge, this is the earliest documented evidence of novel RNA viruses persisting within a host-associated microbiome, extending the observed preservation timescale from centuries to over fourteen millennia. Our findings demonstrate that permafrost is a viable substrate for paleovirological discovery extending beyond the host organism and opening new opportunities for reconstructing ancient microbial and viral ecosystems.