Getting the tree for plotGMPhyloMorphoSpace

[Nadja Pöllath]

Hi all,

I am working with different populations of a single species (sheep). I now want to visualize the (dis)similarities between my populations using plotGMPhyloMorphoSpace. Being a newbie with phylogeny and trees and the like, I feel a bit uncertain, if I am getting this right. I thought of getting the tree from the nj function (ape package) with the LS means distance matrix from my advanced.procD.lm as input. Is this the right way to do this?

Best and many thanks for the great work you're doing,

Nadja

[Mike Collyer]

Hello Nadja,

Your plan is nearly perfectly correct except that instead of the $LS.obs.means.dist output from advanced.procD.lm, use $LS.means.  Verify that the means extracted this way have the same row names as the tips on your phylogeny (but the order is not important).  plotGMPhyloMorphoSpace should be able to do the rest.

Good luck and for a “newbie”, your intuition was excellent!

Best,

Mike

--
You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To post to this group, send email to geomorph-...@googlegroups.com.
Visit this group at https://groups.google.com/group/geomorph-r-package.
To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/3c70460f-b413-4eab-9166-00a8c1779690%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Nadja Pöllath]

Hello Mike,

many thanks for your answer. That was fast!

Probably, a bit off-topic: What happened to the wiki? Esp. your ANOVA lectures are extremely helpful. I tried to reach it a few days ago via the git hub site but was only re-directed to the starting page.

Thanks again and best,
Nadja

[Mike Collyer]

Hi Nadja,

The wiki is no longer available because (1) too many of the examples were antiquated and we do not have the time resources to update them all, (2) we are hoping to move to a vignette system, and (3) we felt that users were relying on them too much and seeking formal training too little.  This thrusts us geomorph maintainers into too much of a pro bono consulting/advising role, which is just a bit excessive, given that geomorph is something we do in addition to our regular jobs.  So to put it simply, we do not have the manpower to maintain the wikis in all regards, so we were forced to abandon them.

Cheers!

Mike

--
You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To post to this group, send email to geomorph-...@googlegroups.com.
Visit this group at https://groups.google.com/group/geomorph-r-package.

To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/3b9ff2bb-7106-4f23-baa9-9862358f4e66%40googlegroups.com.

[Sergio Ferreira Cardoso]

Dear all,

I know that my question is not exactly related to Nadja's one, but since it's on the same function, I will post it on this discussion.

I'm interested in reconstructing ancestral shapes for the internal nodes of my tree. In fact, I did it witht his function, but I was wandering if it is not possible to apply correlation structure different from BM, like Pagel's Lambda or Ornstein-Uhlenbeck. In fact, I guess this is possible in ace. However, this packaged appears to not be ready to deal with multivariate data.
Is it possible to perform such analysis with plotGMPhyloMorphoSpace?

Thank you.

Best regards,
Sérgio.

[Adams, Dean [EEOBS]]

Sérgio,

 

At present, there is no correct way of doing non-BM ancestral state estimation for multivariate data.

 

For multivariate data, all presently proposed methods for evaluating non-Brownian motion evolutionary models (e.g., OU, EB, etc.) display very high rates of misspecification (type I error). That is, they find strong support for more complex models than were used to generate the data (e.g., generate multivariate data under BM, and the methods find strong support for OU). This is undesirable to say the least, and some methods are even worse than this!  A recent review of the problems is found in Adams and Collyer, 2018: Systematic Biology.

 

By extension of these results, working ‘backwards’ to estimate ancestral states using anything except BM will not work as intended.

 

At the present time, if one wishes to make correct inferences with multivariate data, one is restricted to Brownian motion.

 

Dean

 

Dr. Dean C. Adams

Director of Graduate Education, EEB Program

Professor

Department of Ecology, Evolution, and Organismal Biology

       Department of Statistics

Iowa State University

www.public.iastate.edu/~dcadams/

phone: 515-294-3834

--

You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To post to this group, send email to geomorph-...@googlegroups.com.
Visit this group at https://groups.google.com/group/geomorph-r-package.

To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/ea4fb7af-f6dc-4ccc-b78e-d706129f29de%40googlegroups.com.

[Sergio Ferreira Cardoso]

Hi Dean,

Thank you for your answer. I'll read this paper, for sure. I believe it's this one: Adams, Dean C., and Michael L. Collyer. "Multivariate Phylogenetic Comparative Methods: Evaluations, Comparisons, and Recommendations." Systematic biology 67.1 (2017): 14-31.

Thank you once again.

Sérgio.

To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-package+unsub...@googlegroups.com.
To post to this group, send email to geomorph...@googlegroups.com.

[Nadja Pollath]

Dear all,

I was away from GMM for some time and am now starting with the new geomorph. I was able to sort everything out - with the help of the vignettes and the discussions here. But now I am stuck. Where do I get the LS.means now for the nj function?

Best,

Nadja

[Adams, Dean [EEOB]]

Nadja,

 

There are multiple ways to do this, depending on your analytical pipeline.

 

For instance, if you have multiple specimens per group, one could perform a linear model using ‘procD.lm’, followed by ‘pairwise’. In the output from pairwise, the first component of the LS.means contains these  (e.g., PW.output$LS.means[[1]] ). 

 

If you look at the help file for procD.lm there is an example of both the anova and the pairwise tests.

 

This is one of many ways to obtain them.

 

Dean

 

Dr. Dean C. Adams

Director of Graduate Education, EEB Program

Professor

Department of Ecology, Evolution, and Organismal Biology

Iowa State University

https://www.eeob.iastate.edu/faculty/adams/

phone: 515-294-3834

To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/7806f6d5-0c0a-4408-b7ad-1bf3615afb00n%40googlegroups.com.

[nas...@gmx.net]

Dean,

 

thanks - that's it. The output of pairwise was just too long or I was too stupid to find it.

 

Many thanks again,

Nadja

 

Gesendet: Sonntag, 07. März 2021 um 15:22 Uhr
Von: "Adams, Dean [EEOB]" <dca...@iastate.edu>
An: "geomorph-...@googlegroups.com" <geomorph-...@googlegroups.com>
Betreff: RE: [geomorph-r-package] Re: Getting the tree for plotGMPhyloMorphoSpace

You received this message because you are subscribed to a topic in the Google Groups "geomorph R package" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/geomorph-r-package/PVvfk4FRMvQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to geomorph-r-pack...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/geomorph-r-package/CH2PR04MB661749FD113EB40ABF980C78A2949%40CH2PR04MB6617.namprd04.prod.outlook.com.