[R-sig-phylo] diversification rates with two traits present simultaneously

[Elizabeth Christina Miller]

Hello,

I am wondering what comparative method(s) is appropriate for testing if
diversification rates are highest when two traits are present together,
rather than one alone? Specifically, if I have two binary traits, let's say
freshwater/marine and temperate/tropical, what is the best way to test if
diversification rates are highest in tropical+freshwater groups, as opposed
to tropical+marine or temperate+freshwater? I think MuSSE can do this, but
my tree is large so I want to avoid issues associated with rate
heterogeneity. Are there any alternatives?
--
Elizabeth Miller
PhD Candidate: Wiens Lab
Ecology and Evolutionary Biology
University of Arizona

[[alternative HTML version deleted]]

_______________________________________________
R-sig-phylo mailing list - R-sig-***@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-***@r-project.org/

[William Gearty]

I'm not sure of any particular methods that do this, but one preliminary
approach that you could try is combining the two traits into a single trait
(00, 01, 10, and 11) which you could then analyze with any of the standard
methods. You could then look for whether any of the states have higher or
lower diversification rates.

Another approach, particularly easy for large trees, is to calculate the
equal splits metric for your tips (evol.distinct in picante) and then
perform a t-test or other statistical test using your trait data. Come to
think of it, you could actually keep your original two traits and perform
an ANOVA on the equal splits measurements, which would allow you to test
for a significant relationship between either of your traits and the
diversification rate and whether there's a significant interaction between
your two traits which produces higher or lower div. rates.

Best,
Will

On Mon, Jun 18, 2018, 4:46 PM Elizabeth Christina Miller <

Post by Elizabeth Christina Miller
Hello,
I am wondering what comparative method(s) is appropriate for testing if
diversification rates are highest when two traits are present together,
rather than one alone? Specifically, if I have two binary traits, let's say
freshwater/marine and temperate/tropical, what is the best way to test if
diversification rates are highest in tropical+freshwater groups, as opposed
to tropical+marine or temperate+freshwater? I think MuSSE can do this, but
my tree is large so I want to avoid issues associated with rate
heterogeneity. Are there any alternatives?
--
Elizabeth Miller
PhD Candidate: Wiens Lab
Ecology and Evolutionary Biology
University of Arizona
[[alternative HTML version deleted]]
_______________________________________________
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at

[[alternative HTML version deleted]]

_______________________________________________
R-sig-phylo mailing list - R-sig-***@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-***@r-project.org/

[Gilles Benjamin Leduc]

Hello,

Personally, I would advise you not to use a tree, for your 2 traits times 2 (+/-) makes 4 combinations, meaning a tree that is dichotomic cannot reveal anything! Can you quantify your "diversification" without "rate"? How do you obtain your "diversification rates"? If it is something you can count, you may try an old fashion Khi-square test… you would then know if there is a correlation…

Have a nice day,

Benjamin

Post by Elizabeth Christina Miller
Hello,
I am wondering what comparative method(s) is appropriate for testing if
diversification rates are highest when two traits are present together,
rather than one alone? Specifically, if I have two binary traits, let's say
freshwater/marine and temperate/tropical, what is the best way to test if
diversification rates are highest in tropical+freshwater groups, as opposed
to tropical+marine or temperate+freshwater? I think MuSSE can do this, but
my tree is large so I want to avoid issues associated with rate
heterogeneity. Are there any alternatives?
--
Elizabeth Miller
PhD Candidate: Wiens Lab
Ecology and Evolutionary Biology
University of Arizona
[[alternative HTML version deleted]]
_______________________________________________
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo

_______________________________________________
R-sig-phylo mailing list - R-sig-***@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-***@r-project.org/

[Brian O'Meara]

Rich FitzJohn has programmed something of what Will suggested in
diversitree: look at details in ?make.musse.multitrait. I'm not sure if
there was ever a paper published on this. We do a similar thing to Will's
idea in http://rspb.royalsocietypublishing.org/content/283/1830/20152304
(with the added complication of varying over combinations of traits, since
converting six binary traits to a 2^6 state musse model would have been a
bit ambitious given the data size [or, frankly, given any possible sampling
from the tree of life]). However, neither approach deals with hidden
sources of rate heterogeneity. Jeremy Beaulieu and Daniel Caetano are
always adding stuff to hisse that deals with hidden rates (see
preprint on a hidden
biogeography model at
https://www.biorxiv.org/content/early/2018/04/17/302729), but I don't think
anything has been peer-reviewed and published yet for this use case (I
think the core hisse model is also in RevBayes, and that may have added
options making a musse version possible). If your tree is big enough,
separating it into clades and seeing if the pattern (i.e., when converting
00 -> 1, 01 -> 2, 10 -> 3, 11 -> 4 for musse, and you find that state 3
(aka binary states 10) has highest diversification rate) is consistent
across the clades could be a good way of seeing if there are some kinds of
heterogeneity that matter.

Best,
Brian

_______________________________________________________________________
Brian O'Meara, http://www.brianomeara.info, especially Calendar
<http://brianomeara.info/calendars/omeara/>, CV
<http://brianomeara.info/cv/>, and Feedback
<http://brianomeara.info/teaching/feedback/>

Associate Professor, Dept. of Ecology & Evolutionary Biology, UT Knoxville
Associate Head, Dept. of Ecology & Evolutionary Biology, UT Knoxville
Associate Director for Postdoctoral Activities, National Institute for
Mathematical & Biological Synthesis <http://www.nimbios.org> (NIMBioS)

Post by Elizabeth Christina Miller
Hello,
Personally, I would advise you not to use a tree, for your 2 traits times
2 (+/-) makes 4 combinations, meaning a tree that is dichotomic cannot
reveal anything! Can you quantify your "diversification" without "rate"?
How do you obtain your "diversification rates"? If it is something you can
count, you may try an old fashion Khi-square test… you would then know if
there is a correlation…
Have a nice day,
Benjamin
Le Lundi 18 Juin 2018 23:46 GMT, Elizabeth Christina Miller <

Post by Elizabeth Christina Miller
Hello,
I am wondering what comparative method(s) is appropriate for testing if
diversification rates are highest when two traits are present together,
rather than one alone? Specifically, if I have two binary traits, let's

say

Post by Elizabeth Christina Miller
freshwater/marine and temperate/tropical, what is the best way to test if
diversification rates are highest in tropical+freshwater groups, as

opposed

Post by Elizabeth Christina Miller
to tropical+marine or temperate+freshwater? I think MuSSE can do this,

but

Post by Elizabeth Christina Miller
my tree is large so I want to avoid issues associated with rate
heterogeneity. Are there any alternatives?
--
Elizabeth Miller
PhD Candidate: Wiens Lab
Ecology and Evolutionary Biology
University of Arizona
[[alternative HTML version deleted]]
_______________________________________________
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at

_______________________________________________
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at

[[alternative HTML version deleted]]

_______________________________________________
R-sig-phylo mailing list - R-sig-***@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-***@r-project.org/