Table 2.
mammalian traits & features used to build our viral models.
| Category | Mammalian feature | Reason for inclusion |
|---|---|---|
| Phylogeny | Mean phylogenetic distance to known hosts. | Linked to sharing of viruses between mammals4,73,74. |
| Evolutionary distinctiveness | Can correlate negatively with pathogen species richness75. | |
| Taxonomy & domestication | Order & family | Can affect host-pathogen76, particularly viral, associations4. |
| Domestication | Might influence sharing of viruses between host groups. Domesticated mammals and human might share more viruses with each other than related wild species. | |
| Ecological traits | Morphological traits (body mass) | A key feature in terms of metabolism and adaption to environment. |
| Life-history traits (Maximum age, age at sexual maturity, activity cycle, and migration) | Potentially relevant in terms of within-host dynamics of viruses. | |
| Reproductive traits (gestation period length, litters per year, litter size and weaning age) | ||
| Habitat utilisation | Similar habitat utilisation might correlate with contact with similar viruses. | |
| Diet (proportional use of 10 categories) | Similar dietary habit might associate with similar viral assemblage. | |
| Mean ecological distance | Indicates if a potential host species is ecologically close to or distant from the virus’ preferred host range. We based this distance on a generalised form of Gower’s distance matrices77,78 incorporating all ecological traits. | |
| Geo-spatial features | Geographical range (area size) | Might lead to exposure to larger number or more diverse viruses. |
| Climate (mean temperature & precipitation) | Climate has been shown to influence a number of human and domestic mammal pathogens79,80. | |
|
Natural land cover diversity/Agriculture and farming diversity Mammalian biodiversity Urbanisation/human population |
These factors have been found to influence certain categories of host-pathogen associations81. Supplementary Note 3 lists further details of mammalian geo-spatial feature extraction. |
We trained a suite of models for each virus species with two or more known mammalian hosts (n = 556). Each model comprised the below described features (response variable = 1 if the mammal is known to associate with the focal virus species, 0 otherwise – methods section provides further details). Full description of these features, their sources and justification are listed in Supplementary Note 3.