[R-sig-eco] Distance matrix analyses

[Diego Javier Inclan Luna]

Thank you for your kind response,


I am sorry about the lack of details behind my original question, I will
explain now why is important to me to test for the among group
combinations.


Specifically, I have sampled apple orchards and two neighboring seminatural
habitats (forest and grasslands). I am interested in evaluating how the
community composition change across these different habitats. In
particular, I expect that a significant interaction between the nestedness
of two habitats will be an indicative of high spillover between these
habitats. For instance, I have a distance matrix of the “relativized
nestedness” of Podani and Schmera (2011) and I would like to test the
contribution of forest and grassland habitats to the local diversity of
apple orchards. Therefore, I would like to compute the variances in
nestedness across these different habitats. For example, the “adonis”
function in vegan does this work, but my problem is that I cannot test for
the variance among habitats (following the original example it would be the
*ba*, *ca*, and *cb* combinations).


If I specify in “adonis” to do the grouping as *a*, *b* and *c*, it will do
the test on the *aa*, *bb* and *cc* of the matrix. Is it a way to specify
the groupings so that it will subset the distance matrix to test for *ba*,
*ca* and *cb* combinations?



Thank you,



Diego

2014-09-18 17:29 GMT+02:00 Sarah Goslee <sarah.goslee at gmail.com>:

> Test them for what?
>
> If you can formulate your hypotheses more clearly, you might try the
> model matrix approach outlined in
>
> Legendre, P. and M.-J. Fortin. 1989. Spatial pattern and ecological
> analysis. Vegetatio, 80: 107-138.
>
> and subsequent references.
>
> But there might be better ways, depending on what you are trying to
> achieve.
>
> Sarah
>
>
> On Thu, Sep 18, 2014 at 11:17 AM, Diego Javier Inclan Luna
> <diegojavier.inclanluna at studenti.unipd.it> wrote:
> > Hi,
> >
> > I would like to test dissimilarities among different groups in a dist
> > matrix. For instance, if I have a dist matrix with three factors a, b, c,
> > functions like adonis and betadisper in vegan will test the aa, bb and cc
> > combinations. However, I am interested in testing the ba, ca, and cb
> > combinations. How I could test for these combinations?
> >
> >
> > *a  a   a   b   b   b   c   c  c*
> >
> > *a* aa
> > *a* aa aa
> > *a* aa aa aa
> > *b* ba ba ba bb
> > *b* ba ba ba bb bb
> > *b* ba ba ba bb bb bb
> > *c* ca ca ca cb cb cb cc
> > *c* ca ca ca cb cb cb cc cc
> > *c* ca ca ca cb cb cb cc cc cc
> >
> >
> > Thanks!
> > --
> --
> Sarah Goslee
> http://www.functionaldiversity.org
>



-- 
----

Diego J. Inclan

PhD Student

DAFNAE-Entomology, University of Padova

Viale dell'Università 16, 35020, Legnaro (PD) - Italy

webpage: http://www.biodiversity-lorenzomarini.eu/Diego.html

	[[alternative HTML version deleted]]