Currently sampleStates is being sampled from the respective frequencies of demes (infected classes) at the maxSampleTime. If the tips are sampled throughout the tree, then this can introduce a bias. Ideally we should be sampling states at the times of the tfgy solution that correspond to the sample times of the respective tips.
Currently
sampleStates
is being sampled from the respective frequencies of demes (infected classes) at the maxSampleTime. If the tips are sampled throughout the tree, then this can introduce a bias. Ideally we should be sampling states at the times of thetfgy
solution that correspond to the sample times of the respective tips.