Hi Miriam - 

As you said, all of the pPCA axes should have a phylogenetic covariance of 0. 
If you construct an evolutionary vcv matrix from those variables, the matrix 
should have 0's for all non-diagonal elements. If you use that matrix to 
simulate data, the results will then be identical to the case where you 
simulate each PC independently.

So: it doesn't matter, both are the same!

Luke

On Aug 12, 2013, at 4:53 PM, Miriam Santos-Herrera <[email protected]> 
wrote:

> Dear all,
> 
> I am trying to compare the disparities between two clades, the empirical
> values are very different so now I am wondering, is this significantly
> different than a random expectation?
> 
> to test this I performed 1,000 simulations assuming BM. However here comes
> my question: my traits are size corrected variables (obtained through
> "phyl.resid") which afterwards have been transformed using a phylogenetic
> PCA (using "phyl.pca" with correlation matrix). If I understand well
> what phyl.pca does, my axes in the pPCA present a phylogenetic covariance
> 0. Therefore, does it make any sense to simulate using the vcv matrix? or
> instead of that should I simulate each PC independently (each one with its
> rate parameter), and then combine them to calculate the disparities in each
> simulation?
> 
> any help on this would be very appreciated.
> 
> thank you very much.
> 
> 
> Miriam.
> 
>       [[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]/



Luke Harmon
Associate Professor
Biological Sciences
University of Idaho
208-885-0346
[email protected]

_______________________________________________
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