Error message from Tracer when constructing Bayesian skyline plots

[Melissa]

Hello all,

I am attempting to construct a Bayesian skyline plot in Tracer but receive the following error message: "For the linear or exponential change Bayesian skyline model there should be one fewer group size than population size parameters. Either this is a stepwise (constant) model or the wrong parameters were specified. Please try again." I double checked that my XML file specified the EBSP tree prior with the linear model and made certain I loaded the appropriate .trees file. I believe that the group size and population size parameters mentioned in the error message correspond to the demographic.indicators and demographic.popSize parameters in the .log file, of which I have 65 and 66, respectively. This seems to meet the criteria of the error message (one fewer group size than pop size). I suspect I am missing something obvious but have not found it yet. Any advice would be much appreciated!

Melissa

[Matt Shirley]

Hi Melissa,

I may be wrong on this, but as I understand it Tracer cannot actually be used to reconstruct EBSP models, only the BSP and skyride.  Beast is supposed to out put a CSV file that contains the EBSP skyline data for plotting in another graphics program (R or even Excel would likely work.

Though, that being said, I have had an EBSP analysis running for the past 8 days and still have not gotten the CSV file output despite the trees and logfile already being complete.  So....

Good luck!

Matt

[Melissa]

Hi Matt,
Thanks for the information. I will try graphing the .csv file in R. Good luck with your research! Melissa