Hi
I want to perform and RDA analysis using bacterial community data and environmental data as explanatory variables.
I have some variables that are percentages, others are direct measures of physicochemical parameters and other are ratios.
Should I first scale all variables to make them uniform due to the different units and magnitudes? if so, should I use z-score for all or it is a different way for percentages or ratios?
And after scaling (and not before) should I check for normality (which is the best way here?) and transform those variables with a non-normal distribution to get a distribution closer to normal? should I try different transfromations for percentages (like arcsin) or apply log10 transformation to all?
I am lost at this, before doing an RDA or even a biplot PCA I don't know how to proceed. Advise is welcome. Thanks!
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA