[R-sig-phylo] Call for papers: Special Issue featuring linking micro- and macroevolution in Journal of Evolutionary Biology

[Masahito Tsuboi]

Dear colleagues,

I am very pleased to inform you that a submission of manuscripts to a special 
issue (SI) of Journal of Evolutionary Biology “Inferring macroevolutionary 
patterns and processes from microevolutionary mechanisms” has just opened. On 
behalf of guest editors of our SI (Thibault Latrille, Théo Gaboriau and 
myself), here I warmly invite you to submit your or your student’s manuscript 
studying the relationship between microevolution and macroevolution. Themes 
that we aim to cover in our SI include, but not limited to:

• Paleontological data, theory and methods in relation to contemporary data
• Phylogenetic comparative methods and their applications to empirical research 
(note that JEB publishes Methods Article now)
• Dynamic and process of speciation as a bridge between microevolution and 
macroevolution
• Ecological and functional causes of microevolution and macroevolution
• Influence of developmental architecture on evolutionary trajectories across 
scales
• Variability of traits and genomic sequences across individuals, populations 
and species
• Population-genetic and quantitative-genetic parameters in relation to tempo 
and mode of macroevolution
• Phylogenomics and phylogenetics data, theory and methods across time scales

Please see the following link for further details.

https://onlinelibrary.wiley.com/page/journal/14209101/homepage/Journal_of_Evolutionary_Biology_Call_for_Papers.html

To submit your manuscript to our SI, please prompt your submission to JEB in 
the link above, go to "Step 1: Type, Title, & Abstract” page, then choose 
“Inferring macroevolutionary patterns and processes from microevolutionary 
mechanisms” in the dropdown menu of a question “Is this submission for a 
special issue? Either select "No" or the title of the special issue.”

A rough schedule: we plan to close the call on 31st of August and publish the 
special issue in early 2024. Please note that we will have a rolling editorial 
process (submit-review-revise-publish), meaning that successful manuscripts 
will be electronically published as “Early View” articles as soon as editorial 
processes will be completed, and will later be included in the special issue 
when all articles are processed.

Please feel free to share this invitation with your colleagues. Should you have 
any questions, please do not hesitate to contact me at 
masahito.tsu...@biol.lu.se<mailto:masahito.tsu...@biol.lu.se>.

Best wishes,
Masahito Tsuboi

------
Masahito Tsuboi (PhD)

PostDoc
Lund University/University of Oslo


E-mail : masahito.tsu...@biol.lu.se<mailto:masahito.tsu...@biol.lu.se>
Mobile : +46(0)733868631
Website : https://www.masahitotsuboi.com


[[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] Postdoc position on PCM of trait evolution in Stockholm/Sweden

[Masahito Tsuboi]

Dear PCM colleagues,

I am writing on behalf of my colleague John Fitzpatrick regarding a postdoc 
position on phylogenetic comparative study of trait evolution in his lab at 
Stockholm University in Sweden. Job ad is here 
(https://web103.reachmee.com/ext/I007/927/job?site=8=UK=d3e6a58db9058c5eab7ea3e324f063f6_id=19770).

The deadline to apply for this position is 10th of February. If you have 
question,. please contact John (cc’d).

All the best,
Masahito Tsuboi

[[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] phylogenetic correction for multiple samples per species along an ecological gradient

[Masahito Tsuboi]

Hi list,

I second Jon's suggestion for using phylogenetic mixed model. This seems to be 
the closest to what Nicolay is after.

Yet another option is "SLOUCH" package in R, which can handle within-species 
variation while performing phylogenetic corrections. Here are some example 
codes. 

https://kopperud.github.io/slouch/articles/introduction.html 


Hope you will find a suitable solution from one of those.

Best,
Masahito

> On 1 Aug 2018, at 14:37, jonnations  wrote:
> 
> You will want to try Bayesian mixed modeling which can handle your species
> # problem by using “species” as a group level (i.e. “random”) effect. I
> would recommend r packages brms and MCMCglmm and their vignettes. Both have
> detailed “phylo” examples.
> 
> Maybe check out the r-sig-mixed-model listserv as well. Lots of similar
> questions and good advice show up on there.
> 
> Good luck!
> Jon
> -- 
> Jonathan A. Nations
> PhD Candidate
> Esselstyn Lab 
> Museum of Natural Sciences 
> Louisiana State University
> 
>   [[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/

Re: [R-sig-phylo] Interpretation of standard errors of parameter estimates in OUwie models

[Masahito Tsuboi]

Dear R-sig-phylo contributors,

Thanks a lot for sharing your illuminating thoughts on the list.

I agree with points raised by others, and I'd like to add one suggestion. As 
already mentioned by several, alpha and sigma-squared can be hard to estimate 
with certainty. However, in some cases, the stationary variance 
(sigma-squared/2 alpha) converges well while sigma-squared and alpha separately 
do not. What such case means is that the expected amount of variance at an OU 
equilibria (= stationary variances) can be estimated reliably, but how long 
adaptation takes to get there (= alpha, or half life, log(2)/alpha) is highly 
uncertain. I thus would suggest Rafael to compute stationary variances from 
estimated sigma-squared and alpha, and see if you can make a better sense. 

Good luck. Thanks again for putting an interesting question on the table and 
for sharing thoughts.

Best,
Masahito

> On 5 Apr 2018, at 00:41, Marguerite Butler  wrote:
> 
> Aloha Rafael,
> 
> I am very glad that you are thinking about power of your analyses and 
> confidence in your parameter estimates.  Folks have chimed in with many 
> useful responses, however, I believe your original question was whether 
> uncertainty in parameter estimates indicate bad models and limits of 
> interpretation?
> 
> It’s important to note that model fit and parameter estimates are two 
> different aspects of assessment. It is perfectly possible to have robust 
> model selection (decent power) but have parameters that have a lot of 
> uncertainty and are hard to identify with any precision (large confidence 
> intervals).  That was pretty well demonstrated in Cressler, Butler, King 
> 2015. The specific outcome is data and model dependent, but as a 
> generalization, model selection is more robust than parameter estimation, and 
> within the parameters, the position of the thetas are a little bit easier to 
> narrow down than the alphas and sigmas. 
> 
> This is not too surprising when you think about the information content of 
> the data. We are really trying to fit some relatively complex models with few 
> data (one point per species, with phylogenetic structure to boot), and the 
> intuitive interpretation is that there are many ways to the same outcome.  If 
> the outcome is the particular pattern of phenotypes along the tree, it might 
> be possible to get that outcome with strong selection and little 
> stochasticity, or weak selection toward optima very far apart (and the 
> observed phenotypes are essentially “on the way” toward the optima), or very 
> weak selection and strong stochasticity (they were just knocked about that 
> way by drift), etc.  While this last scenario is most different than the 
> first two (the first two are more OU-type, whereas the last is more BM-like 
> in being dominated by stochasticity), my point is that there may not 
> necessarily be a unique set of parameters that explain the data. 
> 
> Nevertheless, it may still be possible to say something biologically useful. 
> First of all, I noticed that you are doing SIMMAP which is great for 
> exploring phylogenetic uncertainty, but this is a bit arms-length with regard 
> to OU parameter and model uncertainty. A bit apples and oranges. I would 
> strongly recommend that you try some parametric bootstrapping on your OU 
> models for both model selection and parameter estimation. In OUCH, there are 
> built-in facilities to make this really easy, probably there are in OUwie as 
> well (sorry I only stick with OUCH :). The key for the model selection 
> bootstrap is that you can simulate the data using the best-fit model, but 
> then fit all of the alternative models to see how often your best model 
> performs. I like to point folks to my former grad student’s paper (Scales et 
> al 2009).  If you want to be even more stastically thorough you can do the 
> Botteinger, Coop, and Ralph 2012 approach which basically does parametric 
> simulations on both the best and alternative models.   The parametric 
> bootstraps on the parameter estimates is much easier. and you can just run 
> the simulations and fit only the best model and draw your CI’s from that. 
> Maybe you’ve done this already. 
> 
> With regard to what can be said, as Liam and Brian and others have said, many 
> times there are large CIs because we encounter flat or ridged likelihood 
> surfaces.  So that may limit your conclusions somewhat to saying something 
> like we don’t know what alpha is, precisely, but we know it’s greater than 
> zero.  Etc.  In your specific case, what you want to know is whether theta 1 
> > theta 2, or something like that, right? So without needing to know the 
> precise location of the theta’s I would bet that you can look through your 
> parametric simulations and see how often that condition is satisfied.  You 
> may find that you don’t know what they are exactly, but one is always greater 
> than the other. 
> 
> I love your study question by 

[R-sig-phylo] Type 2 and 3 SS in pgls

[Masahito Tsuboi]

Dear all,

Hi. My name is Masahito. I am running ANCOVA type model (continuous response 
variable against mix of continuous and categorical independent variables) using 
pglmEstLambda function in CAIC package, and wonder if I can show anova table 
based on type 2 or 3 sum of square, which are not default. 

Thank you very much in advance for any advises.

Sincerely yours,
Masahito

--
Inst Ecology  Genetics/Animal Ecology
Evolutionary Biology Centre (EBC)
Uppsala University
Norbyvägen 18D
75236 Uppsala, Sweden

E-mail  : masahito.tsu...@ebc.uu.se
Phone   : +46(0)184712672
Fax : +46(0)184717725
Mobile  : +46(0)729139735



[[alternative HTML version deleted]]

___
R-sig-phylo mailing list
R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo