[R-sig-phylo] Fwd: rate units in fitContinuous
Thanks, Josef (and those that answered privately). Based on what you wrote, it seems to me that using sigsq as a measure of rate of evolution would make sense for BM, but the interaction between sigsq and alpha in OU would make the interpretation of sigsq as a rate measure complicated, if not unfeasible. Would you agree? On Mon, Dec 7, 2015 at 9:49 PM, Josef C Uyedawrote: > Karla, > The units of sigsq are in your trait units^2 per time unit of your > phylogeny. So if your trait is in cm and your phylogeny in millions of > years, then the units of sigsq are in cm^2/my. > > Comparing BM to OU units is more complicated. Gene Hunt discussed this in > the following paper: > http://paleobiol.geoscienceworld.org/content/38/3/351.short > Long story short, they are in the same units, but it may not be > particularly meaningful to compare them between OU and BM models. For > example, you can have an extremely high sigsq in the OU model, but if alpha > is very high as well, you may end up with very slow rates of evolution when > observed over the lifetime of a phylogeny (specifically, the stationary > variance of the OU process is sigsq^2/(2*alpha), thus if alpha is high > enough the stationary variance will tend towards 0, and all the traits > across taxa will be virtually identical). > > Hope this helps. > Josef Uyeda > > > > On 12/07/2015 03:35 PM, Karla Shikev wrote: > >> Dear all, >> >> I am fitting a BM model using fitContinuous, a vector of trait values and >> a >> time-calibrated tree and then I use the sigsq parameter as my measure of >> the rate of evolution of the trait in question. My questions are: >> >> (1) what is the unit of the sigsq parameter? >> (2) to what extent is the sigsq in a BM model comparable to sigsq in an OU >> model, for instance? are they on the same units? >> >> Thanks! >> >> Karla >> >> [[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/
Re: [R-sig-phylo] Fwd: rate units in fitContinuous
Yes Karla, I think this is clearly stated in the paper to which Josef is referring. Julien De : R-sig-phylo <r-sig-phylo-boun...@r-project.org> de la part de Karla Shikev <karlashi...@gmail.com> Envoyé : mardi 8 décembre 2015 11:57 À : r-sig-phylo@r-project.org Objet : [R-sig-phylo] Fwd: rate units in fitContinuous Thanks, Josef (and those that answered privately). Based on what you wrote, it seems to me that using sigsq as a measure of rate of evolution would make sense for BM, but the interaction between sigsq and alpha in OU would make the interpretation of sigsq as a rate measure complicated, if not unfeasible. Would you agree? On Mon, Dec 7, 2015 at 9:49 PM, Josef C Uyeda <pseudac...@gmail.com> wrote: > Karla, > The units of sigsq are in your trait units^2 per time unit of your > phylogeny. So if your trait is in cm and your phylogeny in millions of > years, then the units of sigsq are in cm^2/my. > > Comparing BM to OU units is more complicated. Gene Hunt discussed this in > the following paper: > http://paleobiol.geoscienceworld.org/content/38/3/351.short > Long story short, they are in the same units, but it may not be > particularly meaningful to compare them between OU and BM models. For > example, you can have an extremely high sigsq in the OU model, but if alpha > is very high as well, you may end up with very slow rates of evolution when > observed over the lifetime of a phylogeny (specifically, the stationary > variance of the OU process is sigsq^2/(2*alpha), thus if alpha is high > enough the stationary variance will tend towards 0, and all the traits > across taxa will be virtually identical). > > Hope this helps. > Josef Uyeda > > > > On 12/07/2015 03:35 PM, Karla Shikev wrote: > >> Dear all, >> >> I am fitting a BM model using fitContinuous, a vector of trait values and >> a >> time-calibrated tree and then I use the sigsq parameter as my measure of >> the rate of evolution of the trait in question. My questions are: >> >> (1) what is the unit of the sigsq parameter? >> (2) to what extent is the sigsq in a BM model comparable to sigsq in an OU >> model, for instance? are they on the same units? >> >> Thanks! >> >> Karla >> >> [[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/ ___ 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/
Re: [R-sig-phylo] Fwd: rate units in fitContinuous
Fantastic. Thanks so much! On Tue, Dec 8, 2015 at 10:19 AM, Julien Clavel <julien.cla...@hotmail.fr> wrote: > Yes Karla, > > I think this is clearly stated in the paper to which Josef is referring. > > Julien > > > De : R-sig-phylo <r-sig-phylo-boun...@r-project.org> de la part de Karla > Shikev <karlashi...@gmail.com> > Envoyé : mardi 8 décembre 2015 11:57 > À : r-sig-phylo@r-project.org > Objet : [R-sig-phylo] Fwd: rate units in fitContinuous > > Thanks, Josef (and those that answered privately). > > Based on what you wrote, it seems to me that using sigsq as a measure of > rate of evolution would make sense for BM, but the interaction between > sigsq > and alpha in OU would make the interpretation of sigsq as a rate measure > complicated, if not unfeasible. Would you agree? > > > > > > On Mon, Dec 7, 2015 at 9:49 PM, Josef C Uyeda <pseudac...@gmail.com> > wrote: > > > Karla, > > The units of sigsq are in your trait units^2 per time unit of your > > phylogeny. So if your trait is in cm and your phylogeny in millions of > > years, then the units of sigsq are in cm^2/my. > > > > Comparing BM to OU units is more complicated. Gene Hunt discussed this in > > the following paper: > > http://paleobiol.geoscienceworld.org/content/38/3/351.short > > Long story short, they are in the same units, but it may not be > > particularly meaningful to compare them between OU and BM models. For > > example, you can have an extremely high sigsq in the OU model, but if > alpha > > is very high as well, you may end up with very slow rates of evolution > when > > observed over the lifetime of a phylogeny (specifically, the stationary > > variance of the OU process is sigsq^2/(2*alpha), thus if alpha is high > > enough the stationary variance will tend towards 0, and all the traits > > across taxa will be virtually identical). > > > > Hope this helps. > > Josef Uyeda > > > > > > > > On 12/07/2015 03:35 PM, Karla Shikev wrote: > > > >> Dear all, > >> > >> I am fitting a BM model using fitContinuous, a vector of trait values > and > >> a > >> time-calibrated tree and then I use the sigsq parameter as my measure of > >> the rate of evolution of the trait in question. My questions are: > >> > >> (1) what is the unit of the sigsq parameter? > >> (2) to what extent is the sigsq in a BM model comparable to sigsq in an > OU > >> model, for instance? are they on the same units? > >> > >> Thanks! > >> > >> Karla > >> > >> [[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/ > [[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/