Individual and temporal variation in pathogen load predicts long-term impacts of an emerging infectious disease

Konstans Wells; Rodrigo K. Hamede; Menna E. Jones; Paul A. Hohenlohe; Andrew Storfer; Hamish I. McCallum

179 DFTD-induced mortality size accounts for tumour size, while tumour growth was 180 modelled as a logistic function with the growth parameter  sampled as an input parameter. 181 We allowed for latency periods  between infection and the onset of tumour growth, which 182 was also sampled as an input parameter. We assumed no recovery from DFTD, which 216 We generated key summary statistics from the case study, in which DFTD was 217 expected to have been introduced shortly before the onset of the study (Hamede et al. 2015) , 218 and a pre-selection of simulation scenarios, in which juveniles never comprised > 50% of the 219 population, DFTD prevalence at end of 10-year-period was between 10 and 70%, and the age 220 of individuals with growing tumours was  52 weeks. Hereafter, we refer to 'prevalence' as Figure S4 ). 268 Inference of input parameters was only possible for some parameters, whereas 95% The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/392324 doi: bioRxiv preprint 12 from field data (Supporting Information Figure S5 , Table S2 ). The posterior of the DFTD- Our results suggest that DFTD will not necessarily cause local Tasmanian devil extinction or 289 even long-term major declines, whereas the extirpation of DFTD or coexistence/endemicity is The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/392324 doi: bioRxiv preprint 13 models, however, differ also from our approach in that they ignore spatial structure and do 299 not account for the uncertainty in unknown parameters such as disease-induced mortality and 300 disease transmission rates.