Dear community,
I am rather new to phylogenetic comparative methods and seek your advice.
I want to compare a continuous trait (dependent) between species belonging to either groupA and groupB (groupA and groupB are two states of a categorical binary trait in this case). Using traditional statistics, a Mann-Whitney U test lets me reject the null hypothesis, as clearly the trait values of groupA tend to be greater than the trait values of groupB (see attached plot).
However, how can I implement a phylogenetic correction into the above analysis? I have already performed a PGLS and calculated phyogenetic signal for the regression using Pagel's lambda in ape/nlme (lambda-0.98). I have found that the traits evolve according to the phylogeny, as after performing a likelihood ratio test I found that the calculated lambda is not different than lambda=1.
My question is how do i go on from that point to test whether the trait values are different between groupA and groupB, taking into account the phylogenetic nonindepednence of the traits?
I understand that if the lambda of the regression would not be significantly different compared to lambda = 0, then I could just treat the trait values of each species as independent data points, but I do not understand how to compare trait values between the two groups if 0
Seems like you are on the right track. Perhaps this will help:
https://cran.r-project.org/web/packages/pez/vignettes/pez-pglmm-overview.pdf
Best, D. Booth
Thank you David this was indeed helpful!
I am working on this and I will keep updating this post based on the progress I make, so it is there for everyone in need
.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R