Estimates of pathogen exposure predict varying transmission likelihood: Host contact and shedding patterns may clarify disease dynamics in desert tortoises Gopherus agassizii

These datasets (S1-S4) document the transmission of a bacterial pathogen (Mycoplasma agassizii) between desert tortoises (Gopherus agassizii). The desert tortoises were experimentally introduced in captivity and were used to create and compare models predicting transmission probability given data on the hosts and their interactions. Datasets S1 & S2 include variables describing the individual tortoises, e.g. id, sex, and variables describing the length of their interaction, e.g., number of days cohabitating, hours of direct contact. Two types of analysis were run on these data (survival models and generalized linear mixed models or GLMMs), which required the data to be structured in different ways: the S1 dataset is structured for use in survival analysis while the S2 dataset was used in GLMMs. The response variable or event of interest in both methods was the infection status of the exposed tortoise after a period of interaction. Infection status was defined in three waysand determined using qPCR of tissue samples collected at intervals the results of which are presented in the S3 dataset. The S2 dataset includes additional variables estimating the infection level (also based on data in S3) of infected tortoises involved in the interaction with the focal host. Interaction time and the amount of bacteria present in an infected host were used to calculate dose variables that represent the intensity of exposure to the pathogen. The analyses allowed us to identify interactions that have high transmission likelihood, and so we explored the contact patterns of a wild tortoise population (25 individuals with overlapping or contiguous home ranges) to estimate how frequently high-risk contacts occur (S4 dataset). This dataset includes all interactions (tortoise ids of interacting pair, date & time interaction began, and interaction duration) documented between tortoises fitted with proximity logging devices. Each device detects other devices when tortoises are approximately 10 cm apart and ends an interaction when tortoises have remained further than 10 cm for 1 minute.