We set up the prior using a lognormal distribution (offset = 63, mean = 10.4, standard deviation = 0.24), such that the mean age was at 73 Ma and the 95% upper boundary was at 78 Ma.
Scenario 3: a very recent divergence time scenario was tested, based on the first appearance of the red fox in the fossil record (0.5-1 Mya) [ 30, 31], setting the minimum age of the root height to 0.5 My and using a lognormal prior.
The cost associated with further investigations was modelled using a lognormal distribution.
HZ and PHN utility decrements were tested using a lognormal distribution [ 13].
Seven calibration points were used as priors for divergence times of certain splits, using a lognormal distribution of prior probability.
We calibrated this node using a lognormal prior (mean in real space = 1, log standard deviation = 1.25, offset = 14).
For all datasets, the analyses were run under a GTR + Γ model for each partition, using a lognormal relaxed clock.
Determinants for serological response at baseline were assessed with logistic regression analysis and for individual serotypes using a lognormal model.
In this analysis, the mean prior for the cytochrome b mutation rate was set 10 times higher than the rate for introns to account for the fact that approximately a 10 times faster rate has been found for mitochondrial sequences with respect to intron sequences [ 35], using a lognormal distribution with a standard deviation of 1.
For all date priors, we used a lognormal distribution with a hard minimum bound set at youngest possible date of first appearance in the fossil record.
To estimate the evolutionary rate we used a lognormal distribution as prior, with a lognormal mean of −6.603 s/s/y and a lognormal stdev of 0.711.
