Unveiling Novel Viral Diversity, Biogeography, and Host Networks in Wildlife Through High-Throughput Sequencing Data Mining.

≈75% of emerging pathogens originating from wildlife. However, viral diversity within wildlife remains insufficiently explored. This work performs an extensive analysis of 57 536 publicly high-throughput sequencing datasets from wild mammals and birds, resulting in the generation of ≈613.45 million assembled contigs, including 131 509 potential viral contigs identified through BLASTn and BLASTx searches. Following the exclusion of index hopping contamination, 9788 are categorized into 25 viral families with known zoonotic potential. These results indicate significant spatial and host-specific variability in viral distribution and reveal a positive correlation between viral diversity and host biodiversity. Rodents, bats, ungulates, and anseriformes exhibit the highest viral diversity. Notably, 50% of the viral sequences exhibit <90% amino acid identity to known viruses, indicating of potential novel viruses. Host-virus network uncovers 458 associations, 67.9% are unreported. Further, sequences of avian influenza viruses are identified in goats, while SARS-CoV-2 are detected in goats, ferrets, porpoises, cactus mice, and house finches. These findings highlight the largely uncharacterized viral diversity in wildlife, underscore the urgent requirement for surveillance at the wildlife-livestock interfaces. Additionally, this work develop the Animal Pathogen Decoding Platform, to facilitate the retrieval and analysis of viral contigs, thereby reducing computational redundancies in future research.