Acceptance proportions too high

[Ziv Lieberman]

Hi all,

My apologies for posting multiple questions in rapid succession (in fact, I may have more to ask very soon!).

I'm currently running a data-free MCMCTree analysis to evaluate effective priors in paml 4.10.9. I'm finding that the first of the five printed acceptance proportions is suitable (~0.3), but the other four are all 1.0 or very close to 1. 

Since I believe that the finetune block in the control file is now deprecated, is there anything to do to ameliorate this? Perhaps is this just expected behavior during a data-free run, and will only be informative when using data?

Thanks for all the help!

-Ziv

[Sishuo Wang]

Dear Ziv,

Would it be possible to share the inputs with us or by email? That would likely help better understand the case.

best,

sishuo

[Ziv Lieberman]

Hi again,

Just an update that when running the analysis with data, the acceptance proportions were all acceptable, so it seems to be something about the data-free run.

-Ziv