Procrustes analysis with multiple datasets

Question

I am interested in community responses to an environmental gradient. I have species data (abundances) for 4 different groups (invertebrates, diatoms, aquatic plants and fish). And I also have a set of environmental descriptors (physico-chemical parameters of the streams and catchments). I have all this information for 10 streams.

I would like to compute a Procrustes analysis to see if the groups respond the same way to the environmental gradient. I ran different RDAs (1 RDA per group) and a PCA on then environmental descriptors. Ideally, I would like to run a Procrustes analyses between my 4 RDAs and 1 PCA (i.e. 5 configurations in total). However, the function procrustes from the vegan package only allows using 2 configurations...

Is there a way to these 5 configurations in the same analysis?

Answer 1

You need generalized Procrustes analysis. It seems to be available at least in the following R packages: shapes, factoMiner, EvoMorph and obviously in several others.

Alternatively you may consider INDSCAL: see R packages smacof, multiway, sensoMiner and others you find searching R package lists.