Unrealistically recent divergence time estimates
Miguel Blázquez
Dear all,
We are trying to time-calibrate the phylogeny of a fungal genus in BEAST v1.10.4 and are getting unrealistically recent divergence time estimates. We are not really sure where the problem is, perhaps you could help us spot it!
Firstly, we generated a fossil-calibrated phylogeny including representatives of the entire Ascomycota phylum to determine the divergence time between our genus and its sister taxa, in a similar way to Mitchell et al. (2021). We obtained a most recent common ancestor with a mean age of 52.89 Mya and 95% HPD at 40.85 and 65.51 Mya, which is similar to the numbers obtained by previous authors.
The problem comes when we attempt to use that age as the secondary calibration point: all the divergence time estimates in the tree turn up to be less than 1 Mya, including the MRCA itself. Indeed, they seem to be expressed as evolutionary distances instead of Mya. So, we are assuming that the issue is somewhere in the xml file of this second tree. This is what we did in BEAUti:
PARTITIONS: we set the partition schemes suggested by PartitionFinder and stablished a single clock model and partition tree.
TAXA: we forced the monophyly of our genus as well as the relationshiphs among the genera that were used as outgroup.
SITES: we set the substitution models suggested by PartitionFinder for each partition.
CLOCKS: we selected a lognormal uncorrelated relaxed clock (the "Use continuous quantile parametrization" box was not checked)
TREES: we used the "Speciation: Birth-Death Process" as tree prior (once the time-calibration issue is resolved we will compare different tree priors) and started with a random tree.
PRIORS: we used the ages obtained in the previous phylogeny to stablish the secondary calibration point. This was at the tmrca() of a monophyletic clade formed by our genus and its sister taxa, and we used a normal prior with mean 52.89 and stdev 7.5.
MCMC: we used a chain of 2x10^8, sampling every 20,000 steps.
We are quite lost, as this is not that different
to what we did in the fossil-calibrated tree. Can you see something strange
that we are missing?
Thank you very much in advance!
Miguel
Literature:
Mitchell, J. K., Garrido-Benavent, I., Quijada, L., Pfister, D. H. (2021). Sareomycetes: more diverse than meets the eye. IMA fungus 12(1): 1-36.
.
