Due to the necessity to control human-to-human spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the overwhelming majority of the generated data on this virus was solely related to the genomic characteristics of strains infecting humans; conversely, this work aimed to recover and analyze the diversity of viral genomes from non-human sources. From a set of 3595 publicly available SARS-CoV-2 genome sequences, 128 lineages were identified in non-human hosts, the majority represented by the variants of concern Delta (n = 1105, 30.7percent-flag-change) and Alpha (n = 466, 12.9percent-flag-change), followed by B.1.1.298 lineage (n = 458, 12.7percent-flag-change). Environment, Neovison vison, Odocoileus virginianus and Felis catus were the non-human sources with the highest number of lineages (14, 12 and 10, respectively). Phylogenomic analyses showed viral clusters from environmental sources, N. vison, O. virginianus, Panthera tigris, and Panthera leo. These clusters were collectively related to human viruses as well as all other non-human sources that were heterogeneously distributed in the phylogenetic tree. Further, the genetic details of viral genomes from bats and pangolins were independently investigated owing to their high divergence, revealing five distinct clusters. Cluster 4 exclusively included bat-sourced genomes and the SARS-CoV-2 reference strain Wuhan-01. In summary, this study identified new genetic landmarks of SARS-CoV-2 evolution. We propose potential interspecies transmission routes of SARS-CoV-2 between animals and humans, which should be considered in order to establish better pathogen surveillance and containment strategies.
