Hi Mercedes, The relaxed bm output summarizes the shifts sampled during the rj-mcmc, so it’s not surprising that it will show some shifts in rate along some branches. The important part of the output is the posterior probability of a shift at a given node — if this is low, then there is little support for the rate change(s) summarized in the plot. So it’s not unexpected to find variable rates estimated, but no overall support for a relaxed bm model over a single bm rate. Be aware though that Tracer uses a harmonic mean estimator of the marginal likelihood when computing Bayes factors and that this approach can sometimes perform poorly, particularly when the difference between the homogeneous and heterogenous rate models are subtle (see Baele et al. 2012)
Best, Graham Ref: Baele G., et al. "Improving the accuracy of demographic and molecular clock model comparison while accommodating phylogenetic uncertainty." Molecular biology and evolution 29.9 (2012): 2157-2167. ------------------------------------------------------------ Graham Slater Peter Buck Post-Doctoral Fellow Department of Paleobiology National Museum of Natural History The Smithsonian Institution [NHB, MRC 121] P.O. Box 37012 (202) 633-1316 slat...@si.edu<mailto:slat...@si.edu> www.fourdimensionalbiology,com On Nov 20, 2013, at 12:21 PM, Mercedes Burns <mercedes.bu...@gmail.com<mailto:mercedes.bu...@gmail.com>> wrote: Dear list, I am investigating a set of continuously varying morphological traits for males and females. I expected some of these traits to have a functional relationship that would translate to their tips states and evolutionary rates of change being correlated. Using PGLS (correcting for body size) and threshold modeling in phytools, I've found this seems to be the case for some but not all of my pairs of traits. I've been using the rjmcmc.bmfunction in geiger to investigate whether a lack of correlation between pairs of rates may be due to rate heterogeneity. I planned to do this by comparing the likelihoods of models of global rate Brownian motion: rjmcmc.bm(tree, trait, summary=TRUE, filebase="traitBM", ngen=50000000, samp=1000, type="bm") and relaxed BM with the reversible jump process-enabled: rjmcmc.bm(tree, trait, summary=TRUE, ngen=50000000, filebase="traitJMPRBM", samp=1000, type="jump-rbm") When I checked these analyses in Tracer, I was surprised to find that there is little difference in log-likelihoods of these models and it seems that even if relaxed local rates are allowed, no rate shifts are actually accepted although I have evidence of inferred jumps in my output. Is this to be expected if global rate Brownian motion is the better fitting model? Is there a problem with having the reversible jump process available? I'm also not sure how I should ultimately compare the log-likelihoods given that the model is also an experimental parameter in the jump-rbm process. I should probably also note that my tree has about 30 tips, which is on the lower end of what Eastman et al. (2011) used in their simulations. Is there a better way to approach this question? I'd appreciate any advice. -- Mercedes Burns Ph.D. Candidate BEES Program Department of Entomology University of Maryland 4112 Plant Sciences Bldg. College Park, MD 20742 (301) 405-7518 [[alternative HTML version deleted]] _______________________________________________ R-sig-phylo mailing list - R-sig-phylo@r-project.org<mailto:R-sig-phylo@r-project.org> https://urldefense.proofpoint.com/v1/url?u=https://stat.ethz.ch/mailman/listinfo/r-sig-phylo&k=diZKtJPqj4jWksRIF4bjkw%3D%3D%0A&r=9yz%2F9T5DbvC96h83J%2BUFBQ%3D%3D%0A&m=L5rIIFDGiwkq5Bb8P7s%2BGSlsZV5FrVFarBFP15q5NIk%3D%0A&s=f551dcf15ea3bfc504047291bff416df8149b7fb29d2c5dfd6848605d91ccb85 Searchable archive at https://urldefense.proofpoint.com/v1/url?u=http://www.mail-archive.com/r-sig-phylo%40r-project.org/&k=diZKtJPqj4jWksRIF4bjkw%3D%3D%0A&r=9yz%2F9T5DbvC96h83J%2BUFBQ%3D%3D%0A&m=L5rIIFDGiwkq5Bb8P7s%2BGSlsZV5FrVFarBFP15q5NIk%3D%0A&s=1918cb9345b82012461085bfb81cb59426375d5abed5d88390a39bb961449dff _______________________________________________ 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/
