Quantifying transmission of emerging zoonoses: Using mathematical models to maximize the value of surveillance data

Monique R. Ambrose; Adam J. Kucharski; Pierre Formenty; Jean-Jacques Muyembe-Tamfum; Anne W. Rimoin; James O. Lloyd-Smith

We conducted a variety of sensitivity analysis tests using simulated datasets to assess 521 how robust the method was over a range of parameter values and assumption violations (full 522 descriptions are provided in S1 Text). The method continued to perform well even at very high 523 spillover rates (S7 Fig, S2 Table) and when the offspring distribution used in simulations 524 differed from the one assumed in the inference (S8 Fig, S3 Table) . In some situations, assuming 525 a larger broader contact zone than the one used for simulations could lead to an overestimation of 526 R and an underestimation of λ z (S4 and S5 Tables). This outcome is consistent with what was 527 observed in the monkeypox analysis where assuming a larger spatial scale for the broader contact 528 zone corresponded to a higher estimate of R and a smaller estimate of the spillover rate (Fig 5) . 529