Estimation of divergence times in phylogenetic trees using sequence data
becomes increasingly popular, but so far dating studies have given widely different results,
and especially datings of the lower nodes within the angiosperms and metazoans, have given
much older ages than those obtained from the fossil record. It has been concluded in different
studies that more taxa, and more fossils are needed for more reliable age estimates. For this
reason, a dating method that can handle very large data sets with multiple fossil constraints is
necessary.
Chronograms obtained by e.g. penalized likelihood and Bayesian methods, often
adds a large "ghost range" to the fossil record, and produces chronograms with a more or less
smooth appearance, even if the corresponding phylograms have apparently very
heterogeneous rates. Compared to the other methods, our recently developed method,
PATHd8, gives the results with the best agreement with the fossil record, which coincides
with the least smooth appearance of the chronograms. When other programs often run into
computational problems when analysing trees with hundreds of leaves, PATHd8 has no
problems analysing thousands of taxa instantaneously. An arbitrary number of fossil age
constraints can be specified, either as fixed-, minimum or maximum age.
With our new method, the biggest problem in dating studies is that we need
more fossils, and these fossils must be well dated and assigned to the correct branches of the
phylogeny. Therefore, to accomplish divergence time estimates, which hopefully approximate
the real ages, biologists now need to cooperate with palaeontologists.