Changed as per a recent Slack discussion. The text of the tutorial is now consistent with the attached scripts in setting the branchRates argument of dnPhyloCTMC() to part_rate[i] rather than part_rate_mult[i]: using the latter would cause the induced prior on branch lengths / tree length to diverge from the user-specified prior. Following @hoehna's recommendation, the concentration parameter of the Dirichlet distribution has been increased to improve convergence.
Changed as per a recent Slack discussion. The text of the tutorial is now consistent with the attached scripts in setting the
branchRatesargument ofdnPhyloCTMC()topart_rate[i]rather thanpart_rate_mult[i]: using the latter would cause the induced prior on branch lengths / tree length to diverge from the user-specified prior. Following @hoehna's recommendation, the concentration parameter of the Dirichlet distribution has been increased to improve convergence.