This function takes a prior function and generates a starting point that can be entered for startpar
in the function bayou.mcmc
overparameterize.startingPoint(prior, tree, dat)A prior function
A phylogenetic tree of class 'phylo'
A named data vector
This function creates an "overparameterized" starting point for running the mcmc. It gives n-1 tips a unique optimum close to the actual data value. This is useful if you expect steep likelihood peaks that may be hard to find, as these often will be easier to access from this overparameterized model. Generally, the overparameterization will have a very high likelihood and a very low prior.