I want to explore correlated ecological traits accounting for phylogeny. I have many continuous and categorical variables and one categorical response variable with three levels. Does anyone knows an adequate test for my data?
Hi Luca,
What kind of analysis do you intend to do? A priori I would recommend BayesTraits: http://www.evolution.reading.ac.uk/BayesTraitsBeta/. You have a good example here: http://nunn.rc.fas.harvard.edu/groups/anthrotree2011/wiki/3507d/Isabella_Capellini__BayesTraits.html. Also you might take a look to this site (https://nunn.rc.fas.harvard.edu/groups/pica/wiki/79203/77_Mixture_of_discrete_and_continuous_data_two_approaches.html. Feel free to email me if you have any doubt.
Cheers.
You can try MacClade. You can map characteristics onto a phylogeny that way.
Thank you Davida, I had a look to MacClade but as far as I could see you can only use binary dependent variables, I have a three level dependent variable
I actually don't know what 454 is, I'm rather inexpert in this kind of analyses, but I would use the phylogenetic tree on mammals from Bininda-Edmond 2007
Try R+ape, http://cran.r-project.org/web/packages/ape/,
In ape you can use the Mantel test to explore correlations between your variables and phylo data.
The vegan package in R is another one you can use if you break your data down into number of species.
Hi Luca,
What kind of analysis do you intend to do? A priori I would recommend BayesTraits: http://www.evolution.reading.ac.uk/BayesTraitsBeta/. You have a good example here: http://nunn.rc.fas.harvard.edu/groups/anthrotree2011/wiki/3507d/Isabella_Capellini__BayesTraits.html. Also you might take a look to this site (https://nunn.rc.fas.harvard.edu/groups/pica/wiki/79203/77_Mixture_of_discrete_and_continuous_data_two_approaches.html. Feel free to email me if you have any doubt.
Cheers.
- Badges
- Science topic