You want to use correlation matrix here not covariance matrix as you want to estimate the phylogenetic variance. There are many different ways to scale a covariance matrix into a correlation matrix (which may match different model of evolution you are assuming). If you use a correlation matrix and you get a phylogenetic variance, which you want to use to get Pagel’s lamda or phylogenetic heritability. You may want to have a look at our recent work - https://ecoevorxiv.org/su4zv/ - see the section “Including and Testing the Phylogenetic Effect” in Discussion
3 Likes