Dear all,

I want to explore gradual/punctual mode of trait evolution, taking into
account phylogenetic uncertainty. Thus, I have 1000 trees  from bayesian
posterior distribution (each tree with ~150 tips). It's very time-consuming
to evaluate kappa in BayesTraits, so I decided to use one of the R
packages. Unlike lambda estimation, kappa drove me to a standstill. So, I
have two questions. I would be extremely grateful for any help or advice
for this problems.

1)
In fitContinuous, for example, there are boundaries on the kappa (min=0,
max=1), although,  according to Pagel, kappa varies from 0 to 3.
Can you suggest, how to overcome this obstacle? A "degree" of
punctuality/graduality  is very critical for my research aims.
I tried to induce boundaries by
bounds=list(kappa = c(min = 0, max = 3))
The highest  value of kappa i received was 1.04   (it's quite
questionably).

2) Next, I'm in doubt about testing received values.
I tried this for one of trees
kappa<-fitContinuous(tree, x, model="kappa",  bounds=list(kappa = c(min =
0, max = 3)))
kappa0<-fitContinuous(kappaTree(tree, 0), x, model="BM")
LRT<-2*(kappa$Trait1$lnl-kappa0$Trait1$lnl)
p.value<-pchisq(LRT, df=1)

Does it make sense?
Is it easier to use another package?

Thank you in advance!


Cheers,
Alice

-- 
Alice O. Vershinina, M.Sc.
Department of Karyosystematics,
Zoological Institute of Russian Academy of Sciences;
Phone: +7(911)758-2794;
e-mail: [email protected]
http://under-the-trees.blogspot.com/ <http://under-the-trees.blogspot.ru/>

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-phylo mailing list - [email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/[email protected]/

Reply via email to