Dear all,
Sorry for shameless self-promotion: I wrote a
plug-in script inBV_bs to help use
MCMCtree to estimate species div time under
complex substitution models that are not currently available in mcmctree like C60, EX2, LG4M, w/ or w/o PMSF approximation. Briefly, it uses a simple bootstrap approach to approximate the Hessian evaluated at the MLE of the branch lengths, which is needed for
mcmctree's approximate likelihood approach to work. Simulations show that, by mcmctree's approxL method, when seq div is large enough e.g., fast-evolving or deep-time, complex models may be a better choice for time and rate estimation [Supplementary Note 4 in Ref1].
Also great thanks to Sandra Álvarez-Carretero and Edmund Moody from University of Bristol who contribute to testing its early versions and to making it much more user-friendly. I believe with a community-driven effort, mcmctree and paml will become even more powerful for researchers in the field.
Any Qs or suggestions are highly appreciated.
best,
sishuo
References