Large variation among different replicates but with same parameters

[XuWeiEvo]

Hi!

I used treePL to analyze my dataset. I ran PRIME to find the optimal parameters and used cross-validation (CV) to determine the best smooth value. However, I ran 10 replicates and noticed a large variation in the time of nodes. Furthermore, removing some fossil calibrations seemed to make the variation bigger among replicates, while removing some priors only resulted in small changes. I am unsure how to proceed with these results. Could you please provide some helpful tips? Thank you so much!

Sincerely,
Wei XU