Nestedness analyses in Vegan
On 05/12/2011, at 11:38 AM, Ivailo wrote:
On Mon, Nov 28, 2011 at 8:32 PM, Ricardo Solar <rrsolar at gmail.com> wrote:
Hi people; I'm thinking about a solution to show that a sample coming from 3 types of killing solutions are nested. However, when I tried to do this with nestedness function in Vegan I couldn't specify the groups. I guess it's impossible at all, so my question is: is there any other analysis where I can show this nestedness pattern? Att, RSolar
Dear Ricardo, as you didn't specify your data structure (and that is important for successfully performing a nestedness analysis), we have to guess it. In in any case it should be entered as "community data" - i.e. site x species incidence matrix. An interesting aspect of your question is how you should organise the "community data" - will you have three separate matrices and compare the nestedness metric among them or will you combine these in one common matrix to study (common) species responses to the three killing solutions. I don't have an answer for that, but perhaps someone on the list might provide further insight...
Ricardo, I think this cannot be done in vegan. In vegan, we make a distinction between community Null models and nestedness statistics (metrics). The Null models generate random communities, and nestedness statistics are evaluated on those. The Null models for random communities are generated with constraints on marginal sums (in presence/absence data species frequencies, species richnesses), but there is no option for having clustered structures for such matrices. If you want to have something like that you must work it out yourself: you should split your matrix to submatrices, generate random communities for each and then again assemble the submatrices together. This is certainly doable, but I don't claim it be easy. Cheers, jari oksanen
Jari Oksanen, Dept Biology, Univ Oulu, 90014 Finland jari.oksanen at oulu.fi, Ph. +358 400 408593, http://cc.oulu.fi/~jarioksa