Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[Liam J. Revell]

I agree with Julien about this. If you have very few observations per 
species in your original data, it might even be useful to just compute a 
single pooled estimate of the within-species variance, and then use that 
to obtain estimates of the standard error for each species mean. To do 
this, first calculate a pooled variance in the normal way 
(https://en.wikipedia.org/wiki/Pooled_variance), and then divide the 
square-root of this variance by the square root of the sample size for 
each species to get an estimate of the standard error for that species. 
This works better if your data have been log-transformed in which case 
an assumption of homogeneity of the variance is more likely to hold.


I also agree with Dave, though, that if the Brownian model doesn't fit 
your data reasonably well, then sigma^2 is pretty difficult to 
interpret. (I hate to point to one of my own papers, but a couple of 
years ago some colleagues & I published a study in which we made a set 
of 'phylogenetic Anscombe' datasets - each of which showed a really 
different among-species pattern, and yet that all had the same value of 
sigma^2: https://dx.doi.org/10./2041-210X.13067. The point of this 
article, like Anscombe 1973, was about graphing - but it may also apply 
to your problem.)


All the best, Liam

Liam J. Revell
University of Massachusetts Boston
Universidad Católica de la Ssma Concepción
web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org

Academic Director UMass Boston Chile Abroad:
https://www.umb.edu/academics/caps/international/biology_chile

On 9/24/2020 1:46 PM, Julien Clavel wrote:

[EXTERNAL SENDER]

Hi Karla,

As pointed out by Liam you can account for the sampling error with 
fitContnuous. Pagel's lambda corresponds to BM + noise at the tips (e.g. 
sampling error). If after accounting for this nuisance (for a set of 
competitive models) you find that BM is better than the other models, then it 
will makes sense to consider the sigsq as a reasonable measure of rate even if 
your Pagel's lambda estimate was low.

Best wishes,

Julien

De : R-sig-phylo  de la part de Karla Shikev 

Envoyé : jeudi 24 septembre 2020 03:41
À : Liam J. Revell 
Cc : R Sig Phylo Listserv ; Theodore Garland 

Objet : Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic 
signal

Thanks, Liam! That was really helpful. But what would be an alternative to
estimate evolutionary rates in those cases? Simply the sigsq estimate under
a white noise model? (which I assume is simply the variance in the original
data?)

Karla

On Wed, Sep 23, 2020 at 10:23 PM Liam J. Revell  wrote:


Hi Karla.

   > I just wanted to make sure that the estimate of sigma^2 in a BM model
   > in fitContinuous is a meaningful measure of the rate of evolution of
   > the trait even if its phylogenetic signal is low.

I would say that the answer to this question (if it were posed as one)
is 'no.' Another way to put this is that sigma^2 is probably not a
meaningful measure of the rate of evolution for your character if
Brownian motion is a (very) poor fit to your data.

It's kind of equivalent to saying 'is the slope of a fitted linear
regression a good measure of the rate of change in y as a function of x
if a linear model is a very poor fit to my data?' (That is, you can
still find a value of the slope that maximizes the probability of your
data under a linear model, and this slope accurately give the mean
effect of x on y over the range of your observations for x - but it may
not mean a whole lot about how x & y genuinely co-vary.)

All the best, Liam

Liam J. Revell
University of Massachusetts Boston
Universidad Católica de la Ssma Concepción
web: http://faculty.umb.edu/liam.revell/, 
https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.phytools.org%2Fdata=02%7C01%7Cliam.revell%40umb.edu%7Cc32f6ae954334ee27bbf08d860a9554a%7Cb97188711ee94425953c1ace1373eb38%7C0%7C0%7C637365627682724545sdata=mPJpNrUUHdUZ7nGRvCSc0yJSGpec32VK8%2F5ijH3ATTw%3Dreserved=0

Academic Director UMass Boston Chile Abroad:
https://www.umb.edu/academics/caps/international/biology_chile

On 9/23/2020 9:36 PM, Karla Shikev wrote:

**[EXTERNAL SENDER]

Dear All,

Thanks for your fantastic feedback. Just a bit more background: I'm
interested both in the phylogenetic signal and the evolutionary rate. I
just wanted to make sure that the estimate of sigma^2 in a BM model in
fitContinuous is a meaningful measure of the rate of evolution of the
trait even if its phylogenetic signal is low.

thanks again!

Karla

On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland
mailto:theodore.garl...@ucr.edu>> wrote:

  Carla, I can also suggest looking at some of the papers on this,
  such as:

  Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
  phylogenetic signal in comparative data: behavioral traits are more
  labile. Evolution 57:717–745.


  Ives, A. R., P. E. Midford, and T. Garland, J

Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[Julien Clavel]

Hi Karla,

As pointed out by Liam you can account for the sampling error with 
fitContnuous. Pagel's lambda corresponds to BM + noise at the tips (e.g. 
sampling error). If after accounting for this nuisance (for a set of 
competitive models) you find that BM is better than the other models, then it 
will makes sense to consider the sigsq as a reasonable measure of rate even if 
your Pagel's lambda estimate was low.

Best wishes,

Julien

De : R-sig-phylo  de la part de Karla Shikev 

Envoyé : jeudi 24 septembre 2020 03:41
À : Liam J. Revell 
Cc : R Sig Phylo Listserv ; Theodore Garland 

Objet : Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic 
signal 
 
Thanks, Liam! That was really helpful. But what would be an alternative to
estimate evolutionary rates in those cases? Simply the sigsq estimate under
a white noise model? (which I assume is simply the variance in the original
data?)

Karla

On Wed, Sep 23, 2020 at 10:23 PM Liam J. Revell  wrote:

> Hi Karla.
>
>  > I just wanted to make sure that the estimate of sigma^2 in a BM model
>  > in fitContinuous is a meaningful measure of the rate of evolution of
>  > the trait even if its phylogenetic signal is low.
>
> I would say that the answer to this question (if it were posed as one)
> is 'no.' Another way to put this is that sigma^2 is probably not a
> meaningful measure of the rate of evolution for your character if
> Brownian motion is a (very) poor fit to your data.
>
> It's kind of equivalent to saying 'is the slope of a fitted linear
> regression a good measure of the rate of change in y as a function of x
> if a linear model is a very poor fit to my data?' (That is, you can
> still find a value of the slope that maximizes the probability of your
> data under a linear model, and this slope accurately give the mean
> effect of x on y over the range of your observations for x - but it may
> not mean a whole lot about how x & y genuinely co-vary.)
>
> All the best, Liam
>
> Liam J. Revell
> University of Massachusetts Boston
> Universidad Católica de la Ssma Concepción
> web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org
>
> Academic Director UMass Boston Chile Abroad:
> https://www.umb.edu/academics/caps/international/biology_chile
>
> On 9/23/2020 9:36 PM, Karla Shikev wrote:
> > **[EXTERNAL SENDER]
> >
> > Dear All,
> >
> > Thanks for your fantastic feedback. Just a bit more background: I'm
> > interested both in the phylogenetic signal and the evolutionary rate. I
> > just wanted to make sure that the estimate of sigma^2 in a BM model in
> > fitContinuous is a meaningful measure of the rate of evolution of the
> > trait even if its phylogenetic signal is low.
> >
> > thanks again!
> >
> > Karla
> >
> > On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland
> > mailto:theodore.garl...@ucr.edu>> wrote:
> >
> > Carla, I can also suggest looking at some of the papers on this,
> > such as:
> >
> > Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
> > phylogenetic signal in comparative data: behavioral traits are more
> > labile. Evolution 57:717–745.
> >
> >
> > Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
> > variation and measurement error in phylogenetic comparative methods.
> > Systematic Biology 56:252–270.
> >
> >
> > Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
> > evolutionary process, and rate. Systematic Biology 57:591–601.
> >
> >
> > Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal
> > with measurement error: a comparison of Mantel tests, Blomberg et
> > al.’s K, and phylogenetic distograms. Evolution 66:2614–2621.
> >
> >
> > Cheers,
> >
> > Ted Garland
> >
> >
> >
> >
> > On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell  > <mailto:liam.rev...@umb.edu>> wrote:
> >
> > Dear Karla.
> >
> > What do you want to use fitContinuous to do?
> >
> > fitContinuous fits multiple models, some of which will result in
> > low
> > values of lambda in a test of phylogenetic signal. In fact, this
> is
> > really the main point of fitContinuous: to compare alternative
> > models
> > for trait evolution. If you just want to estimate the
> > most-likely rate
> > of evolution assuming a Brownian evolutionary process, you could
> > just
> > compute the mean squared PIC value (this is the REML estimate 

Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[David Bapst]

Karla et al.,

There isn't, necessarily, a good alternative. If phylogenetic signal
is very poor, our ability to estimate rates is extremely hindered.
There's no rate of evolution in a white noise model -- values are
simply assigned independently of phylogeny.

One thought I had is that you might be able to at least constrain a
hard *minimum* estimate of the rates of evolution by looking at pairs
of sister species. But that may not be very useful depending on what
question is motivating you to measure evolutionary rate. It may not
even be meaningful in terms of capturing any sort of realistic
estimate of a real world process.

The problem here is that rate, in and of itself, is part of our
assumptions about the process being gradual and accruing variance over
time. If change was instead entirely limited to short punctuations,
and punctuations were randomly distributed in time, then asking how
much change there was over some interval of time (i.e. the rate) would
be rather nonsensical. You could 'measure' the rate but it wouldn't
have anything to do with the process generating the data. Perhaps the
thing to consider is, why do you think this data has such poor
phylogenetic signal to begin with?

In addition to Ted's reading list, these concepts were particularly
clarified for me by Gene Hunt's 2012 paper in Paleobiology on the
inseparability of measures of rate from our assumptions about mode:

https://www.cambridge.org/core/journals/paleobiology/article/measuring-rates-of-phenotypic-evolution-and-the-inseparability-of-tempo-and-mode/2906747B67D16EF002DC8B12CCF96B15

I hope this helps,
-Dave B

On Wed, Sep 23, 2020 at 8:42 PM Karla Shikev  wrote:
>
> Thanks, Liam! That was really helpful. But what would be an alternative to
> estimate evolutionary rates in those cases? Simply the sigsq estimate under
> a white noise model? (which I assume is simply the variance in the original
> data?)
>
> Karla
>
> On Wed, Sep 23, 2020 at 10:23 PM Liam J. Revell  wrote:
>
> > Hi Karla.
> >
> >  > I just wanted to make sure that the estimate of sigma^2 in a BM model
> >  > in fitContinuous is a meaningful measure of the rate of evolution of
> >  > the trait even if its phylogenetic signal is low.
> >
> > I would say that the answer to this question (if it were posed as one)
> > is 'no.' Another way to put this is that sigma^2 is probably not a
> > meaningful measure of the rate of evolution for your character if
> > Brownian motion is a (very) poor fit to your data.
> >
> > It's kind of equivalent to saying 'is the slope of a fitted linear
> > regression a good measure of the rate of change in y as a function of x
> > if a linear model is a very poor fit to my data?' (That is, you can
> > still find a value of the slope that maximizes the probability of your
> > data under a linear model, and this slope accurately give the mean
> > effect of x on y over the range of your observations for x - but it may
> > not mean a whole lot about how x & y genuinely co-vary.)
> >
> > All the best, Liam
> >
> > Liam J. Revell
> > University of Massachusetts Boston
> > Universidad Católica de la Ssma Concepción
> > web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org
> >
> > Academic Director UMass Boston Chile Abroad:
> > https://www.umb.edu/academics/caps/international/biology_chile
> >
> > On 9/23/2020 9:36 PM, Karla Shikev wrote:
> > > **[EXTERNAL SENDER]
> > >
> > > Dear All,
> > >
> > > Thanks for your fantastic feedback. Just a bit more background: I'm
> > > interested both in the phylogenetic signal and the evolutionary rate. I
> > > just wanted to make sure that the estimate of sigma^2 in a BM model in
> > > fitContinuous is a meaningful measure of the rate of evolution of the
> > > trait even if its phylogenetic signal is low.
> > >
> > > thanks again!
> > >
> > > Karla
> > >
> > > On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland
> > > mailto:theodore.garl...@ucr.edu>> wrote:
> > >
> > > Carla, I can also suggest looking at some of the papers on this,
> > > such as:
> > >
> > > Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
> > > phylogenetic signal in comparative data: behavioral traits are more
> > > labile. Evolution 57:717–745.
> > >
> > >
> > > Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
> > > variation and measurement error in phylogenetic comparative methods.
> > > Systematic Biology 56:252–270.
> > >
> > >
> > > Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
> > > evolutionary process, and rate. Systematic Biology 57:591–601.
> > >
> > >
> > > Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal
> > > with measurement error: a comparison of Mantel tests, Blomberg et
> > > al.’s K, and phylogenetic distograms. Evolution 66:2614–2621.
> > >
> > >
> > > Cheers,
> > >
> > > Ted Garland
> > >
> > >
> > >
> > >
> > > On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell  > > 

Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[Karla Shikev]

Thanks, Liam! That was really helpful. But what would be an alternative to
estimate evolutionary rates in those cases? Simply the sigsq estimate under
a white noise model? (which I assume is simply the variance in the original
data?)

Karla

On Wed, Sep 23, 2020 at 10:23 PM Liam J. Revell  wrote:

> Hi Karla.
>
>  > I just wanted to make sure that the estimate of sigma^2 in a BM model
>  > in fitContinuous is a meaningful measure of the rate of evolution of
>  > the trait even if its phylogenetic signal is low.
>
> I would say that the answer to this question (if it were posed as one)
> is 'no.' Another way to put this is that sigma^2 is probably not a
> meaningful measure of the rate of evolution for your character if
> Brownian motion is a (very) poor fit to your data.
>
> It's kind of equivalent to saying 'is the slope of a fitted linear
> regression a good measure of the rate of change in y as a function of x
> if a linear model is a very poor fit to my data?' (That is, you can
> still find a value of the slope that maximizes the probability of your
> data under a linear model, and this slope accurately give the mean
> effect of x on y over the range of your observations for x - but it may
> not mean a whole lot about how x & y genuinely co-vary.)
>
> All the best, Liam
>
> Liam J. Revell
> University of Massachusetts Boston
> Universidad Católica de la Ssma Concepción
> web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org
>
> Academic Director UMass Boston Chile Abroad:
> https://www.umb.edu/academics/caps/international/biology_chile
>
> On 9/23/2020 9:36 PM, Karla Shikev wrote:
> > **[EXTERNAL SENDER]
> >
> > Dear All,
> >
> > Thanks for your fantastic feedback. Just a bit more background: I'm
> > interested both in the phylogenetic signal and the evolutionary rate. I
> > just wanted to make sure that the estimate of sigma^2 in a BM model in
> > fitContinuous is a meaningful measure of the rate of evolution of the
> > trait even if its phylogenetic signal is low.
> >
> > thanks again!
> >
> > Karla
> >
> > On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland
> > mailto:theodore.garl...@ucr.edu>> wrote:
> >
> > Carla, I can also suggest looking at some of the papers on this,
> > such as:
> >
> > Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
> > phylogenetic signal in comparative data: behavioral traits are more
> > labile. Evolution 57:717–745.
> >
> >
> > Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
> > variation and measurement error in phylogenetic comparative methods.
> > Systematic Biology 56:252–270.
> >
> >
> > Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
> > evolutionary process, and rate. Systematic Biology 57:591–601.
> >
> >
> > Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal
> > with measurement error: a comparison of Mantel tests, Blomberg et
> > al.’s K, and phylogenetic distograms. Evolution 66:2614–2621.
> >
> >
> > Cheers,
> >
> > Ted Garland
> >
> >
> >
> >
> > On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell  > > wrote:
> >
> > Dear Karla.
> >
> > What do you want to use fitContinuous to do?
> >
> > fitContinuous fits multiple models, some of which will result in
> > low
> > values of lambda in a test of phylogenetic signal. In fact, this
> is
> > really the main point of fitContinuous: to compare alternative
> > models
> > for trait evolution. If you just want to estimate the
> > most-likely rate
> > of evolution assuming a Brownian evolutionary process, you could
> > just
> > compute the mean squared PIC value (this is the REML estimate of
> > sigma^2).
> >
> > fitContinuous can also take into account sampling error in the
> > estimation of species means, which can be a cause for low
> measured
> > values of phylogenetic signal.
> >
> > Hope this is of some help Karla.
> >
> > All the best, Liam
> >
> > Liam J. Revell
> > University of Massachusetts Boston
> > Universidad Católica de la Ssma Concepción
> > web: http://faculty.umb.edu/liam.revell/,
> > http://www.phytools.org
> > <
> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.phytools.org%2F=02%7C01%7Cliam.revell%40umb.edu%7C4b3649722f5c4a9bb0f508d86021d687%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365045732071994=L7HXU2Ljrj89dZ5ql%2FRWMB7ddrReMWnZxdufzJQ2Nvs%3D=0
> >
> >
> > Academic Director UMass Boston Chile Abroad:
> > https://www.umb.edu/academics/caps/international/biology_chile
> >
> > On 9/23/2020 8:21 PM, Karla Shikev wrote:
> >  > [EXTERNAL SENDER]
> >  >
> >  > Dear friends,
> >  >
> >  > I guess this is as much a philosophical question as it is
> > 

Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[Liam J. Revell]

Hi Karla.

> I just wanted to make sure that the estimate of sigma^2 in a BM model
> in fitContinuous is a meaningful measure of the rate of evolution of
> the trait even if its phylogenetic signal is low.

I would say that the answer to this question (if it were posed as one) 
is 'no.' Another way to put this is that sigma^2 is probably not a 
meaningful measure of the rate of evolution for your character if 
Brownian motion is a (very) poor fit to your data.


It's kind of equivalent to saying 'is the slope of a fitted linear 
regression a good measure of the rate of change in y as a function of x 
if a linear model is a very poor fit to my data?' (That is, you can 
still find a value of the slope that maximizes the probability of your 
data under a linear model, and this slope accurately give the mean 
effect of x on y over the range of your observations for x - but it may 
not mean a whole lot about how x & y genuinely co-vary.)


All the best, Liam

Liam J. Revell
University of Massachusetts Boston
Universidad Católica de la Ssma Concepción
web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org

Academic Director UMass Boston Chile Abroad:
https://www.umb.edu/academics/caps/international/biology_chile

On 9/23/2020 9:36 PM, Karla Shikev wrote:

**[EXTERNAL SENDER]

Dear All,

Thanks for your fantastic feedback. Just a bit more background: I'm 
interested both in the phylogenetic signal and the evolutionary rate. I 
just wanted to make sure that the estimate of sigma^2 in a BM model in 
fitContinuous is a meaningful measure of the rate of evolution of the 
trait even if its phylogenetic signal is low.


thanks again!

Karla

On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland 
mailto:theodore.garl...@ucr.edu>> wrote:


Carla, I can also suggest looking at some of the papers on this,
such as:

Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
phylogenetic signal in comparative data: behavioral traits are more
labile. Evolution 57:717–745.


Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
variation and measurement error in phylogenetic comparative methods.
Systematic Biology 56:252–270.


Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
evolutionary process, and rate. Systematic Biology 57:591–601.


Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal
with measurement error: a comparison of Mantel tests, Blomberg et
al.’s K, and phylogenetic distograms. Evolution 66:2614–2621.


Cheers,

Ted Garland




On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell mailto:liam.rev...@umb.edu>> wrote:

Dear Karla.

What do you want to use fitContinuous to do?

fitContinuous fits multiple models, some of which will result in
low
values of lambda in a test of phylogenetic signal. In fact, this is
really the main point of fitContinuous: to compare alternative
models
for trait evolution. If you just want to estimate the
most-likely rate
of evolution assuming a Brownian evolutionary process, you could
just
compute the mean squared PIC value (this is the REML estimate of
sigma^2).

fitContinuous can also take into account sampling error in the
estimation of species means, which can be a cause for low measured
values of phylogenetic signal.

Hope this is of some help Karla.

All the best, Liam

Liam J. Revell
University of Massachusetts Boston
Universidad Católica de la Ssma Concepción
web: http://faculty.umb.edu/liam.revell/,
http://www.phytools.org



Academic Director UMass Boston Chile Abroad:
https://www.umb.edu/academics/caps/international/biology_chile

On 9/23/2020 8:21 PM, Karla Shikev wrote:
 > [EXTERNAL SENDER]
 >
 > Dear friends,
 >
 > I guess this is as much a philosophical question as it is
methodological. I
 > want to estimate the rate of evolution of a trait (e.g. using the
 > fitContinuous function), but the trait has low phylogenetic
signal (e.g.
 > lambda close to 0). Does it make sense to use fitContinuous
on a trait
 > evolving like that?
 >
 > Thanks for your input.
 >
 > Karla
 >
 >          [[alternative HTML version deleted]]
 >
 > ___
 > R-sig-phylo mailing list - R-sig-phylo@r-project.org

 >

Re: [R-sig-phylo] evolutionary rate for species with low phylogenetic signal

[Karla Shikev]

Dear All,

Thanks for your fantastic feedback. Just a bit more background: I'm
interested both in the phylogenetic signal and the evolutionary rate. I
just wanted to make sure that the estimate of sigma^2 in a BM model in
fitContinuous is a meaningful measure of the rate of evolution of the trait
even if its phylogenetic signal is low.

thanks again!

Karla

On Wed, Sep 23, 2020 at 8:45 PM Theodore Garland 
wrote:

> Carla, I can also suggest looking at some of the papers on this, such as:
>
> Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
> phylogenetic signal in comparative data: behavioral traits are more labile.
> Evolution 57:717–745.
>
>
> Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
> variation and measurement error in phylogenetic comparative methods.
> Systematic Biology 56:252–270.
>
>
> Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
> evolutionary process, and rate. Systematic Biology 57:591–601.
>
>
> Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal with
> measurement error: a comparison of Mantel tests, Blomberg et al.’s K, and
> phylogenetic distograms. Evolution 66:2614–2621.
>
>
> Cheers,
>
> Ted Garland
>
>
>
>
> On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell 
> wrote:
>
>> Dear Karla.
>>
>> What do you want to use fitContinuous to do?
>>
>> fitContinuous fits multiple models, some of which will result in low
>> values of lambda in a test of phylogenetic signal. In fact, this is
>> really the main point of fitContinuous: to compare alternative models
>> for trait evolution. If you just want to estimate the most-likely rate
>> of evolution assuming a Brownian evolutionary process, you could just
>> compute the mean squared PIC value (this is the REML estimate of sigma^2).
>>
>> fitContinuous can also take into account sampling error in the
>> estimation of species means, which can be a cause for low measured
>> values of phylogenetic signal.
>>
>> Hope this is of some help Karla.
>>
>> All the best, Liam
>>
>> Liam J. Revell
>> University of Massachusetts Boston
>> Universidad Católica de la Ssma Concepción
>> web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org
>>
>> Academic Director UMass Boston Chile Abroad:
>> https://www.umb.edu/academics/caps/international/biology_chile
>>
>> On 9/23/2020 8:21 PM, Karla Shikev wrote:
>> > [EXTERNAL SENDER]
>> >
>> > Dear friends,
>> >
>> > I guess this is as much a philosophical question as it is
>> methodological. I
>> > want to estimate the rate of evolution of a trait (e.g. using the
>> > fitContinuous function), but the trait has low phylogenetic signal (e.g.
>> > lambda close to 0). Does it make sense to use fitContinuous on a trait
>> > evolving like that?
>> >
>> > Thanks for your input.
>> >
>> > Karla
>> >
>> >  [[alternative HTML version deleted]]
>> >
>> > ___
>> > R-sig-phylo mailing list - R-sig-phylo@r-project.org
>> >
>> https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylodata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=4u7Q9SP0M9b0l07P13q6zruhfqyMW0AEVIqa%2FT9p1Ko%3Dreserved=0
>> > Searchable archive at
>> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2Fdata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=sKb4MeniYVOo5pQSEUtWRIpz%2BAGHt6wccZ%2B5mPQl7qU%3Dreserved=0
>> >
>>
>> ___
>> 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] evolutionary rate for species with low phylogenetic signal

[Theodore Garland]

Carla, I can also suggest looking at some of the papers on this, such as:

Blomberg, S. P., T. Garland, Jr., and A. R. Ives. 2003. Testing for
phylogenetic signal in comparative data: behavioral traits are more labile.
Evolution 57:717–745.


Ives, A. R., P. E. Midford, and T. Garland, Jr. 2007. Within-species
variation and measurement error in phylogenetic comparative methods.
Systematic Biology 56:252–270.


Revell, L., L. Harmon, and D. Collar. 2008. Phylogenetic signal,
evolutionary process, and rate. Systematic Biology 57:591–601.


Hardy, O. J., and S. Pavoine. 2012. Assessing phylogenetic signal with
measurement error: a comparison of Mantel tests, Blomberg et al.’s K, and
phylogenetic distograms. Evolution 66:2614–2621.


Cheers,

Ted Garland




On Wed, Sep 23, 2020 at 4:33 PM Liam J. Revell  wrote:

> Dear Karla.
>
> What do you want to use fitContinuous to do?
>
> fitContinuous fits multiple models, some of which will result in low
> values of lambda in a test of phylogenetic signal. In fact, this is
> really the main point of fitContinuous: to compare alternative models
> for trait evolution. If you just want to estimate the most-likely rate
> of evolution assuming a Brownian evolutionary process, you could just
> compute the mean squared PIC value (this is the REML estimate of sigma^2).
>
> fitContinuous can also take into account sampling error in the
> estimation of species means, which can be a cause for low measured
> values of phylogenetic signal.
>
> Hope this is of some help Karla.
>
> All the best, Liam
>
> Liam J. Revell
> University of Massachusetts Boston
> Universidad Católica de la Ssma Concepción
> web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org
>
> Academic Director UMass Boston Chile Abroad:
> https://www.umb.edu/academics/caps/international/biology_chile
>
> On 9/23/2020 8:21 PM, Karla Shikev wrote:
> > [EXTERNAL SENDER]
> >
> > Dear friends,
> >
> > I guess this is as much a philosophical question as it is
> methodological. I
> > want to estimate the rate of evolution of a trait (e.g. using the
> > fitContinuous function), but the trait has low phylogenetic signal (e.g.
> > lambda close to 0). Does it make sense to use fitContinuous on a trait
> > evolving like that?
> >
> > Thanks for your input.
> >
> > Karla
> >
> >  [[alternative HTML version deleted]]
> >
> > ___
> > R-sig-phylo mailing list - R-sig-phylo@r-project.org
> >
> https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylodata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=4u7Q9SP0M9b0l07P13q6zruhfqyMW0AEVIqa%2FT9p1Ko%3Dreserved=0
> > Searchable archive at
> https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2Fdata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=sKb4MeniYVOo5pQSEUtWRIpz%2BAGHt6wccZ%2B5mPQl7qU%3Dreserved=0
> >
>
> ___
> 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] evolutionary rate for species with low phylogenetic signal

[Liam J. Revell]

Dear Karla.

What do you want to use fitContinuous to do?

fitContinuous fits multiple models, some of which will result in low 
values of lambda in a test of phylogenetic signal. In fact, this is 
really the main point of fitContinuous: to compare alternative models 
for trait evolution. If you just want to estimate the most-likely rate 
of evolution assuming a Brownian evolutionary process, you could just 
compute the mean squared PIC value (this is the REML estimate of sigma^2).


fitContinuous can also take into account sampling error in the 
estimation of species means, which can be a cause for low measured 
values of phylogenetic signal.


Hope this is of some help Karla.

All the best, Liam

Liam J. Revell
University of Massachusetts Boston
Universidad Católica de la Ssma Concepción
web: http://faculty.umb.edu/liam.revell/, http://www.phytools.org

Academic Director UMass Boston Chile Abroad:
https://www.umb.edu/academics/caps/international/biology_chile

On 9/23/2020 8:21 PM, Karla Shikev wrote:

[EXTERNAL SENDER]

Dear friends,

I guess this is as much a philosophical question as it is methodological. I
want to estimate the rate of evolution of a trait (e.g. using the
fitContinuous function), but the trait has low phylogenetic signal (e.g.
lambda close to 0). Does it make sense to use fitContinuous on a trait
evolving like that?

Thanks for your input.

Karla

 [[alternative HTML version deleted]]

___
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-phylodata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=4u7Q9SP0M9b0l07P13q6zruhfqyMW0AEVIqa%2FT9p1Ko%3Dreserved=0
Searchable archive at 
https://nam10.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-archive.com%2Fr-sig-phylo%40r-project.org%2Fdata=02%7C01%7Cliam.revell%40umb.edu%7C238f2a4ce92f4b62411f08d86017813f%7Cb97188711ee94425953c1ace1373eb38%7C0%7C1%7C637365001363567653sdata=sKb4MeniYVOo5pQSEUtWRIpz%2BAGHt6wccZ%2B5mPQl7qU%3Dreserved=0



___
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] evolutionary rate for species with low phylogenetic signal

[Karla Shikev]

Dear friends,

I guess this is as much a philosophical question as it is methodological. I
want to estimate the rate of evolution of a trait (e.g. using the
fitContinuous function), but the trait has low phylogenetic signal (e.g.
lambda close to 0). Does it make sense to use fitContinuous on a trait
evolving like that?

Thanks for your input.

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/