Statistics with quantitative variables

Good morning:
I have a question about the statistical design of an experiment. I have studied the colonization of microorganisms on different surfaces in different environmental conditions. I have done 16S metabarcoding of these biofilms and I have already done the taxonomic assignment, as the alpha and beta diversity. Now I have studied the surface properties of these materials (roughness and hidrophobicity). The problem is that these variables are quantitative and I have no idea how to statistically interrelate them with my study. Would you know of any idea or methodology? Thank you very much

Hi @gaudy93,

For your alpha diversity, you can just feed it into your favorite correlation/multivariate model. (Which may be why everyone loves alpha diversity). So, a spearman correlation or even, if you see normal residuals, a linear regression.

For beta diversity, adonis will let you pass in continuous or categorical variables, as will bioenv, if it’s an option.

And, as far as features go, this would be a good chance to look at some of the options in Gneiss. The original paper looks at a pH gradient, so it may be applicable to you here.


1 Like