bayou: bayou

This version of bayou has been overhauled to allow more flexible implementation of customized models. Major changes include:

Changed the way MCMC's are run. Users now create an MCMC object which has elements $run and $load, simplifying file management and allowing interrupts & restarts in the MCMC chain without losing past analyses. Custom models can include multiple reversible-jump parameters that shift together on the tree. Custom models can be specified by a customized providing a "model object" which is a list that can be made manually, or using the function makeBayouModel. Predictors with missing data can be imputed assuming Brownian Motion. Phylogenetic heatmaps where branches reflect a particular parameter value averaged over the MCMC can be visualized using plotBranchHeatMap