Hi Gene, Michael and Chris Thanks for your helpful suggestions. Lots of food for thought and plenty to keep me busy over the xmas break!
All the best Joe Dr Joseph Keating Palaeobiology Research Group University of Bristol Bristol Life Sciences Building 24 Tyndall Avenue Bristol BS8 1TQ ________________________________ From: R-sig-phylo <r-sig-phylo-boun...@r-project.org> on behalf of Chris Organ <or...@montana.edu> Sent: 03 December 2024 16:49 To: r-sig-phylo@r-project.org <r-sig-phylo@r-project.org> Subject: Re: [R-sig-phylo] Trait-dependence for an extinct clade Hi Joe, You might also look here for inspiration: https://www.journals.uchicago.edu/doi/10.1086/588076 This "simple test" is flexible and can by adapted for a variety of hypotheses, including those about extinct clades. Best wishes, Chris ______________________________________________________ Chris Organ Deep Time Biology Lab Department of Earth Sciences 202 & 204 Traphagen Hall Montana State University Bozeman, Montana, 59717 or...@montana.edu | https://chrisorgan.github.io | 406.994.6920 On 12/3/2024 5:54 AM, Michael Zyphur via R-sig-phylo wrote: > Hi Joe > > Entirely extinct clades sound like a challenge. As you note, a 'quasse' > approach won’t work for fossil-only datasets because it requires > ultrametric trees. For testing trait-dependent speciation and extinction in > a fossil-only context, a Bayesian framework might be more flexible. > RevBayes allows for custom model specifications and could accommodate the > kinds of trait-dependent processes you’re interested in, even with > non-ultrametric trees. That said, it would involve some setup and possibly > steep learning curves. > > Gene mentions PyRate, which sounds good particularly if you’re open to > using a non-phylogenetic approach. PyRate can analyze stratigraphic range > data and incorporate covariates like body size into its models of > speciation and extinction rates. Although it's Python based, its outputs > are intuitive and if you’re working with stratigraphic range data it could > be super helpful. > > Another option might be integrating phylogenetic and stratigraphic > approaches. For example, if you hypothesize that small-bodied ostracoderms > experience higher rates of extinction, you could test this by examining > whether body size correlates with stratigraphic range lengths. This > wouldn’t capture trait-dependent speciation, but it still might be useful. > Survival analysis or even simple regressions could work here, but PyRate > again could provide a more sophisticated framework if needed. For survival > analysis, Instats has an upcoming seminar on this in case you're > interested: > https://instats.org/seminar/survival-analysis-for-social-and-health-2 > > One thing to keep in mind is how uncertainties in body size estimates, > especially those derived from 3D models, might propagate through your > analyses. Incorporating sensitivity analyses or Bayesian approaches that > explicitly model these uncertainties could help strengthen your results. > Obviously, working with priors and hyperpriors can be troubling with > Bayesian approaches, but they allow substantial degrees of flexibility and > can help estimate parameters that otherwise aren't estimable given the > information in a dataset. > > Hope this is useful. Also, if anyone would be interested in offering an > Instats seminar -- either free of fee-based -- please reach out and I'd be > happy to chat further. Instats is a platform built by and for academic > researchers and we never charge any fees to use our platform to create and > offer any content. > > > Best wishes > > Michael Zyphur > Director > Institute for Statistical and Data Science > *instats.org* <http://instats.org> > > > On Tue, 3 Dec 2024 at 23:37, Hunt, Gene via R-sig-phylo < > r-sig-phylo@r-project.org> wrote: > >> Hi Joe, >> >> The fossil version of BAMM can handle extinct clades and shifts in rate >> regime on branches, but I am not sure if the developers have implemented >> trait-dependent speciation and extinction. You can check out this paper >> and maybe follow-up with the developers: >> >> >> https://academic.oup.com/sysbio/article-abstract/68/1/1/4999317?redirectedFrom=fulltext >> >> Pyrate may be able to do all you need. There are a lot of papers that use >> this tool, and I don’t have the details at hand, but check out the Google >> Scholar page of its developer, Daniele Silvestro: >> >> https://scholar.google.com/citations?user=X1jlzMoAAAAJ&hl=en&oi=ao >> >> Pyrate is a python program (no R interface), but Daniele is known to be >> helpful to users if they have questions. >> >> If the taxa have stratigraphic ranges (rather than all being restricted to >> a single interval), you can use non-phylogenetic approaches to look for an >> association between stratigraphic range and body size. If your hypothesis >> about extinction is correct, large-bodied taxa should have longer >> stratigraphic ranges. The simplest approach would be correlation or >> regression but you can switch to Pyrate (which also run on strat range >> data, rather than a phylogeny) or Capture-Mark-Recapture if you want to >> have a more flexible modeling framework. Those latter approaches can >> handle, for example, temporal or taxonomic variation in preservation and/or >> extinction. I can send references for these approaches to you separately if >> you’d like. >> >> The non-phylogenetic approaches are less suitable for trait-dependent >> speciation, however, because you have to make assumptions about transitions >> between traits (e.g., that small bodied taxa always arise from other >> small-bodied taxa). >> >> Best, >> Gene >> >> Gene Hunt (he/him) >> Curator >> Department of Paleobiology >> w 202.633.1331 hu...@si.edu<mailto:hu...@si.edu> web< >> https://naturalhistory.si.edu/staff/gene-hunt> >> >> SMITHSONIAN INSTITUTION >> NATIONAL MUSEUM OF NATURAL HISTORY >> Facebook<https://www.facebook.com/nmnh.fanpage/> | Twitter< >> https://twitter.com/NMNH> | Instagram< >> https://www.instagram.com/smithsoniannmnh/> >> >> >> From: R-sig-phylo <r-sig-phylo-boun...@r-project.org> on behalf of Joseph >> Keating via R-sig-phylo <r-sig-phylo@r-project.org> >> Date: Tuesday, December 3, 2024 at 3:21 AM >> To: r-sig-phylo@r-project.org <r-sig-phylo@r-project.org> >> Subject: [R-sig-phylo] Trait-dependence for an extinct clade >> External Email - Exercise Caution >> >> Dear All >> >> I�m hoping you can advise me on an appropriate test of trait-dependence >> for an entirely extinct clade. >> >> I want to look at the evolution of body size in ostracoderms - the >> palaeozoic armoured jawless fish. A number of ostracoderm clades show an >> apparent increase in body size through the Devonian. This is taken as a >> classic example of 'Cope's Rule' and has also been linked to competition >> with / predation by jawed vertebrates. >> >> I have some dated trees for a couple of the most diverse ostracoderm >> clades. Each tree consists of around 100 taxa (this might not sound like a >> lot, but is actually around 25% of the total known diversity). I've also >> got a keen MSc student ready to collect body sizes for these groups, >> potentially using some nice 3D models to get volume estimates. >> >> I had initially thought to use �quasse� to model body-size dependent >> speciation and extinction rates, but after some digging it seems like this >> only works for ultrametric trees. Is there a nice statical way of testing >> whether bodysize is correlated with extinction or speciation given a tree >> comprising 100% fossils? My hypothesis is that extinction is more likely in >> ostracoderms with small body sizes. >> >> Any help you can provide would be hugely appreciated! >> >> All the best >> >> Joe Keating >> >> >> >> Dr Joseph Keating >> Palaeobiology Research Group >> University of Bristol >> Bristol Life Sciences Building >> 24 Tyndall Avenue >> Bristol >> BS8 1TQ >> >> [[alternative HTML version deleted]] >> >> [[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/ >> > > [[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/ _______________________________________________ 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/ [[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/
