Species-level predictors?

[fbkarsdorp]

Hi! Thanks for this cool package! Maybe I missed something, but would it also be possible to define features on the level of species in addition to, for example, environmental and spatial features?

[MaximilianPi]

Hi @fbkarsdorp, no we don't support this but we could theoretically implement it.

Best, Max

[fbkarsdorp]

Thanks!

[florianhartig]

In relation to the other post, just to re-iterate on this: @fbkarsdorp, what "features" are you thinking about, and what do you want to model with them?

I would guess that you are speaking about species traits, of which phylogeny might be considered a subclass, is that correct?

The second question is what model structures would you like to see implemented using species traits? There are many options, e.g. trait / phylogeny-structured REs where species that are closer in trait space are closer in their RE values, but there are also other options ...

If you want to test if there is phylogenetic or trait signal, you can do this also on the residuals.

[fbkarsdorp]

Hi @florianhartig, I was planning to apply this on non-biological data, but to entities equivalent of species. So yes, the ability to incorporate specific species traits would be great. Maybe there are better and more efficient ways to do this, but what I hoped to find was a way to compare the impact/influence of say environmental, and spatial factors on the co-occurrence of species with that of specific traits.