question about ESS value

[Zou Bei]

Hi lovely ppl!

 i've made 2 indenpendent run of then datesite and all ESS values are >200. However, after i combined them, there are 2 parameters are shows ESS <200. Can anyone tell me why and do i still able to use this result for my phylo study?

Combined log

2nd log

1st log

[Dylan O'Hearn]

Hi,

I'm not an expert but a basic reason this can happen is that in the two replicate analyses, that parameter converged on different values.  So in each analysis on its own, the ESS is high because it's sampling from the same distribution throughout the chain.  But then when you combine them, it turns out they're different so the ESS becomes low. 

I think as long as the combined ESS is over 200, it should be fine, so you can just run each chain longer (or run a third and fourth maybe, and combine all of them) to bring it up. 

However, something funny is that in the 1st log, freqparameter1 is 0.253 and freqparameter2 is 0.235.  Then in the 2nd log, those are reversed, with 1 bring 0.235 and 2 bring 0.253.  Someone more familiar with the workings of BEAST would have to say for sure, but it kind of seems like maybe they got switched somehow, like just the names are wrong but the actual parameters are the same in both logs.  Just a thought.