Aloha Jacob, Liam, Brian, To add to Brian's and Liam's excellent comments - just remember that any type of ancestral state reconstruction of any form is just a weighted average of the information that you put into the problem. If you have no fossil data or data back in time, it's a weighted average of the tips dictated by the model and/or the algorithm that you use. So point one is keep that in mind.
Second, along the lines of Brian's comments, it is always worth looking at the model outputs - first of all, are the estimates different? If not, then there is no reason to model average, the solution is robust to the model. If they are different, then would it make more sense to interpret each model results one by one and think about what the scenarios imply? It might be useful to consider the most distinct alternative hypotheses, for example. Of course I always only include in the model set the ones that have a biological interpretation (the most robust tests have the most distinct but biologically interpretable hypotheses). It is a way of bracketing the interpretation of an answer. Often model averaging will result in parameter estimates that are intermediate in value - but if none of the models returned that estimate, then what is it exactly? It's a point that is best fit for nothing. But the solution along an entire tree is a solution that is part of a set (a single evolutionary scenario). Mixing and matching from different solutions may not make sense. So I've personally never found a situation where model averaging is useful and sometimes can lead to non-critical thinking. But if you find something, I'm all ears! Best of luck, Marguerite On Thu, Aug 8, 2019 at 10:14 PM Jacob Berv <jakeberv.r.sig.ph...@gmail.com> wrote: > Cool- thanks for the great comments Liam and Brian, as always! > J > > > On Aug 7, 2019, at 11:55 AM, Liam Revell <liam.rev...@umb.edu> wrote: > > > > Hi Jake. > > > > [Edit: I see just now that Brian has also responded to this inquiry. I > > have no doubt that his message is more insightful than mine - but I'll > > nonetheless send what I was writing anyway just in case it contributes > > anything useful to the discussion.] > > > > If you're simply interested in the states at nodes, you might consider > > just multiplying the marginal reconstructions under maximum likelihood > > by the Akaike weights of each model & summing them. > > > > The former are probabilities that the node is in each state conditioned > > on a model, and the latter are the probabilities that each of the models > > is the best of the set. If the models in the set genuinely comprise all > > possible ways in which your character could have evolved (they don't, > > but still), then the model weighted average marginal reconstructions > > should give the total probability that each node is in each state. > > > > If you really want to do stochastic mapping, you might consider some > > kind of rjMCMC in which you sample models & transition rates from their > > joint posterior distribution and then generate stochastic character maps > > based on this sample. This is not implemented in phytools::make.simmap > > (it does MCMC, but only given a model), but is not to hard to envision > > doing, so long as you are careful about designing the rjMCMC. > > > > Now to read Brian's answer.... > > > > All the best, Liam > > > > Liam J. Revell > > Associate Professor, University of Massachusetts Boston > > Profesor Asistente, Universidad Católica de la Ssma Concepción > > web: http://faculty.umb.edu/liam.revell/ < > http://faculty.umb.edu/liam.revell/>, http://www.phytools.org < > http://www.phytools.org/> > > > > Academic Director UMass Boston Chile Abroad (starting 2019): > > https://www.umb.edu/academics/caps/international/biology_chile < > https://www.umb.edu/academics/caps/international/biology_chile> > > > > On 8/7/2019 11:32 AM, Jacob Berv wrote: > >> Dear R-sig phylo > >> > >> I’ve been running a few discrete character Mk type models using > phytool's SIMMAP — and I had the idea that it might be useful to try model > averaging across posterior probabilities for node states. > >> > >> Might this make sense to do, to avoid problems associated with model > ranking via AIC? Ie, average the node state probabilities based on AIC > weights? Is there some fundamental problem with this? > >> > >> I could imagine generating some code to generate all possible > transition models given a set of N states, and then rather than ranking > with AIC, model averaging for parameter estimates (though now that I think > about it, not sure how one might reasonably average a symmetrical rate and > an asymmetrical rate). > >> > >> Best, > >> Jake > >> > >> [[alternative HTML version deleted]] > >> > >> _______________________________________________ > >> R-sig-phylo mailing list - R-sig-phylo@r-project.org <mailto: > R-sig-phylo@r-project.org> > >> > https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylo&data=02%7C01%7Cliam.revell%40umb.edu%7Ce85015cb8c314c684bdc08d71b4c8fee%7Cb97188711ee94425953c1ace1373eb38%7C0%7C0%7C637007887976647805&sdata=F8qdVoT17J%2BuC6wgvFktdI1id%2Bf37hDLe1W17z6OhTw%3D&reserved=0 > < > https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylo&data=02%7C01%7Cliam.revell%40umb.edu%7Ce85015cb8c314c684bdc08d71b4c8fee%7Cb97188711ee94425953c1ace1373eb38%7C0%7C0%7C637007887976647805&sdata=F8qdVoT17J%2BuC6wgvFktdI1id%2Bf37hDLe1W17z6OhTw%3D&reserved=0 > > > >> Searchable archive at > https://nam01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2F&data=02%7C01%7Cliam.revell%40umb.edu%7Ce85015cb8c314c684bdc08d71b4c8fee%7Cb97188711ee94425953c1ace1373eb38%7C0%7C0%7C637007887976657803&sdata=ogCaep2DN92bk6%2FZwOOcpmHaCPQwX7BzNgKo2UMcWtw%3D&reserved=0 > < > https://nam01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2F&data=02%7C01%7Cliam.revell%40umb.edu%7Ce85015cb8c314c684bdc08d71b4c8fee%7Cb97188711ee94425953c1ace1373eb38%7C0%7C0%7C637007887976657803&sdata=ogCaep2DN92bk6%2FZwOOcpmHaCPQwX7BzNgKo2UMcWtw%3D&reserved=0 > > > >> > > > [[alternative HTML version deleted]] > > _______________________________________________ > R-sig-phylo mailing list - R-sig-phylo@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo > Searchable archive at > http://www.mail-archive.com/r-sig-phylo@r-project.org/ > -- ____________________________________________ Marguerite A. Butler Professor Department of Biology 2538 McCarthy Mall, Edmondson Hall 216 Honolulu, HI 96822 Office: 808-956-4713 Dept: 808-956-8617 Lab: 808-956-5867 FAX: 808-956-4745 http://butlerlab.org http://manoa.hawaii.edu/biology/people/marguerite-butler http://www2.hawaii.edu/~mbutler [[alternative HTML version deleted]] _______________________________________________ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
