BEASTX advice on bacterial anlysis

[Nicolás Guerra]

Hi all,

I ran a BEAST 2 analysis on 853 Klebsiella pneumoniae isolates for a time-scaled phylogeny that I plan to use with SPREAD4 later. After a 10% burn-in, parameters like rootHeight, likelihood, and treeLength have ESS values under 100 in Tracer. I don’t have the .state file, so I can't resume the run.

Would the correct strategy be to run new independent replicates with the same XML and then combine the results in LogCombiner? Also, any advice on improving mixing for bacterial datasets would be greatly appreciated (e.g., operator tuning, clock model/prior choices, tree prior, logging frequency).
Thanks!