Re: [R-sig-phylo] OU for non-ultrametric trees

2018-12-16 Thread Marguerite Butler 
Aloha Danielle,

There is no reason at all (mathematically) why you canʻt run
non-ultrametric trees. ouch has ever been limited this way in its
functions, and it also does not require bifurcating trees.

Our original comment was that in most cases people will want to interpret
their results in units of time, which is implied by an ultra metric tree.
However, if there is some other metric that is more useful for the problem
at hand, such as units of mutational change, that is fine too. This may be
more relevant for virus evolution. The important thing is that the branch
lengths have some meaning, and users should put some thought into the
interpretation of the model output.

Marguerite

On Sun, Dec 16, 2018 at 3:09 AM Danielle Miller 
wrote:

> First, I would like to thank all for your replies.
>
> Second, I have a tree constructed of many viral genome sequences. I use
> midpoint rooting in order to run the model on a rooted tree (not using
> outgroup) and do not have three ways splits.
> I’ve probably read this in a paper mentioning that in order to fit OU one
> needs an ultrametric tree (pre 2014).
>
> Looking at the models equations I did not fully understand way this is a
> valid assumption, and wanted to make sure I’m not missing anything
> important.
>
> Once again, many thanks,
> Danelle
>
> > On 12 Dec 2018, at 23:18, David Bapst  wrote:
> >
> > Danielle,
> >
> > Apologies for the late comment on this, but what do you mean by
> non-rooted? Do you mean the tree is rooted, but has a three-way split with
> the outgroup at the root node? You shouldn't do OU on a tree you can't
> assign a root to.
> >
> > I'm also just a bit curious what gave you the sense that OU was
> inappropriate for non-ultrametric trees. To my knowledge, this was only an
> issue in one case, as highlighted by Slater (2014), that one particularly
> (popular) algorithm for fitting evolutionary models doesn't work for OU on
> non-ultrametric trees. I don't think this is as much of a concern in 2018,
> as I think most/all packages that were doing such have fixed their code so
> an alternative algorithm is used when a non-ultrametric tree is given (as
> with a tree containing fossil taxa).
> >
> > Cheers,
> > -Dave
> >
> >
> > On Thu, Dec 6, 2018 at 6:35 AM Danielle Miller  > wrote:
> > Hi all,
> >
> > I have a non-rooted non-ultrametric tree and a corresponding set of a
> single trait values for each one of the tips.
> > I’m interested whether it follows a BM model or an OU.
> >
> > Reading previous comments in the archive, I understood that running an
> OU process is inadequate in a case of non-ultrametric trees, however I did
> not fully understand why.
> > As I use ouch package, what are the consequences of running an unrooted
> tree? And a non-ultrametric one?
> >
> > What would be the best way assessing this issue? Diversitree should be
> more reliable?
> >
> > I will appreciate any explanation and help,
> > Danielle Miller
> > ___
> > R-sig-phylo mailing list - R-sig-phylo@r-project.org  R-sig-phylo@r-project.org>
> > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo <
> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo>
> > Searchable archive at
> http://www.mail-archive.com/r-sig-phylo@r-project.org/ <
> http://www.mail-archive.com/r-sig-phylo@r-project.org/>
> >
> >
> > --
> > David W. Bapst, PhD
> > Asst Research Professor, Geology & Geophysics, Texas A & M University
> > Postdoc, Ecology & Evolutionary Biology, Univ of Tenn Knoxville
> > https://github.com/dwbapst/paleotree <
> https://github.com/dwbapst/paleotree>
>
>
> [[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/
>


-- 

Marguerite A. Butler
Professor

Department of Biology
2538 McCarthy Mall, Edmondson Hall 216
Honolulu, HI 96822

Office: 808-956-4713
Dept: 808-956-8617
Lab:  808-956-5867
FAX:   808-956-4745
http://butlerlab.org
http://manoa.hawaii.edu/biology/people/marguerite-butler
http://www2.hawaii.edu/~mbutler

[[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] OU for non-ultrametric trees

2018-12-16 Thread Danielle Miller 
First, I would like to thank all for your replies. 

Second, I have a tree constructed of many viral genome sequences. I use 
midpoint rooting in order to run the model on a rooted tree (not using 
outgroup) and do not have three ways splits.
I’ve probably read this in a paper mentioning that in order to fit OU one needs 
an ultrametric tree (pre 2014).

Looking at the models equations I did not fully understand way this is a valid 
assumption, and wanted to make sure I’m not missing anything important.

Once again, many thanks,
Danelle 

> On 12 Dec 2018, at 23:18, David Bapst  wrote:
> 
> Danielle,
> 
> Apologies for the late comment on this, but what do you mean by non-rooted? 
> Do you mean the tree is rooted, but has a three-way split with the outgroup 
> at the root node? You shouldn't do OU on a tree you can't assign a root to.
> 
> I'm also just a bit curious what gave you the sense that OU was inappropriate 
> for non-ultrametric trees. To my knowledge, this was only an issue in one 
> case, as highlighted by Slater (2014), that one particularly (popular) 
> algorithm for fitting evolutionary models doesn't work for OU on 
> non-ultrametric trees. I don't think this is as much of a concern in 2018, as 
> I think most/all packages that were doing such have fixed their code so an 
> alternative algorithm is used when a non-ultrametric tree is given (as with a 
> tree containing fossil taxa).
> 
> Cheers,
> -Dave
> 
> 
> On Thu, Dec 6, 2018 at 6:35 AM Danielle Miller  > wrote:
> Hi all,
> 
> I have a non-rooted non-ultrametric tree and a corresponding set of a single 
> trait values for each one of the tips. 
> I’m interested whether it follows a BM model or an OU.
> 
> Reading previous comments in the archive, I understood that running an OU 
> process is inadequate in a case of non-ultrametric trees, however I did not 
> fully understand why. 
> As I use ouch package, what are the consequences of running an unrooted tree? 
> And a non-ultrametric one? 
> 
> What would be the best way assessing this issue? Diversitree should be more 
> reliable?
> 
> I will appreciate any explanation and help,
> Danielle Miller 
> ___
> 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/ 
> 
> 
> 
> -- 
> David W. Bapst, PhD
> Asst Research Professor, Geology & Geophysics, Texas A & M University
> Postdoc, Ecology & Evolutionary Biology, Univ of Tenn Knoxville
> https://github.com/dwbapst/paleotree 


[[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] OU for non-ultrametric trees

2018-12-12 Thread David Bapst 
Danielle,

Apologies for the late comment on this, but what do you mean by non-rooted?
Do you mean the tree is rooted, but has a three-way split with the outgroup
at the root node? You shouldn't do OU on a tree you can't assign a root to.

I'm also just a bit curious what gave you the sense that OU was
inappropriate for non-ultrametric trees. To my knowledge, this was only an
issue in one case, as highlighted by Slater (2014), that one particularly
(popular) algorithm for fitting evolutionary models doesn't work for OU on
non-ultrametric trees. I don't think this is as much of a concern in 2018,
as I think most/all packages that were doing such have fixed their code so
an alternative algorithm is used when a non-ultrametric tree is given (as
with a tree containing fossil taxa).

Cheers,
-Dave


On Thu, Dec 6, 2018 at 6:35 AM Danielle Miller 
wrote:

> Hi all,
>
> I have a non-rooted non-ultrametric tree and a corresponding set of a
> single trait values for each one of the tips.
> I’m interested whether it follows a BM model or an OU.
>
> Reading previous comments in the archive, I understood that running an OU
> process is inadequate in a case of non-ultrametric trees, however I did not
> fully understand why.
> As I use ouch package, what are the consequences of running an unrooted
> tree? And a non-ultrametric one?
>
> What would be the best way assessing this issue? Diversitree should be
> more reliable?
>
> I will appreciate any explanation and help,
> Danielle Miller
> ___
> 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/
>


-- 
David W. Bapst, PhD
Asst Research Professor, Geology & Geophysics, Texas A & M University
Postdoc, Ecology & Evolutionary Biology, Univ of Tenn Knoxville
https://github.com/dwbapst/paleotree

[[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] OU for non-ultrametric trees

2018-12-06 Thread Brian O'Meara 
You can run in OUwie if you set the root.age (we should probably make that
automatic, but haven't yet). See
https://www.rdocumentation.org/packages/OUwie/versions/1.50/topics/OUwie. I
believe the past issue with non-ultrametric trees comes from functions that
handles the creation of the VCV matrix inappropriately by assuming an
ultrametric tree. Using non-ultrametric trees may be possible in other OU
packages (ouch, slouch, phylolm, MVmorph, etc.) but I don't know for sure.

It's also worth checking for why your tree is non-ultrametric. If it has a
mixture of extinct and modern taxa, that's cool. If it's because you think
that the branch lengths reflect the amount of change that'd drive the
continuous character change (i.e., they mean number of generations for taxa
with different generation times, so they don't have the same root to tip
length) that's ok, too. If it's just that those are raw branch lengths
(parsimony, likelihood) without making the tree a chronogram, that's bad --
the main thing OU/BM models are doing is effectively stretching and
smooshing branch lengths (for OU, also adjusting the expected means) so if
the starting branch lengths don't mean something that's relevant, the
parameters you get from stretching also don't mean anything. [I imagine
you're doing it well, just a teachable moment for new students on the list].

Best,
Brian

___
Brian O'Meara, http://www.brianomeara.info, especially Calendar
, CV
, and Feedback


Associate Professor, Dept. of Ecology & Evolutionary Biology, UT Knoxville
Associate Head, Dept. of Ecology & Evolutionary Biology, UT Knoxville



On Thu, Dec 6, 2018 at 7:35 AM Danielle Miller 
wrote:

> Hi all,
>
> I have a non-rooted non-ultrametric tree and a corresponding set of a
> single trait values for each one of the tips.
> I’m interested whether it follows a BM model or an OU.
>
> Reading previous comments in the archive, I understood that running an OU
> process is inadequate in a case of non-ultrametric trees, however I did not
> fully understand why.
> As I use ouch package, what are the consequences of running an unrooted
> tree? And a non-ultrametric one?
>
> What would be the best way assessing this issue? Diversitree should be
> more reliable?
>
> I will appreciate any explanation and help,
> Danielle Miller
> ___
> 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] OU for non-ultrametric trees

2018-12-06 Thread Danielle Miller 
Hi all,

I have a non-rooted non-ultrametric tree and a corresponding set of a single 
trait values for each one of the tips. 
I’m interested whether it follows a BM model or an OU.

Reading previous comments in the archive, I understood that running an OU 
process is inadequate in a case of non-ultrametric trees, however I did not 
fully understand why. 
As I use ouch package, what are the consequences of running an unrooted tree? 
And a non-ultrametric one? 

What would be the best way assessing this issue? Diversitree should be more 
reliable?

I will appreciate any explanation and help,
Danielle Miller 
___
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/