PCoA vs NMDS

Question

[valentinavan]

Hi,

I need to perform some comparative analyses with metagenomics data (abundances). I would like to know whether, in term of final result, using PCoA or NMDS really makes a big difference, and if yes how?

My understanding is that both start with a dissimilarity matrix, they both let you chose the most appropriate distance index but NMDS, contrary to PCoA, performs a series of iterations to adjust the configuration of the data points in the ordinated space until it best fits (matches) the original configuration of the point (which based on their reciprocal distances in the dissimilarity matrix).

1. The final results is the same though, right?
2. Based on what I have to chose PCoA or NMDS?

Thanks in advance

Answer 1

[valentinavan]

I have found good answers:

1. https://www.davidzeleny.net/anadat-r/doku.php/en:pcoa_nmds

2. https://www.researchgate.net/post/How-to-choose-ordination-method-such-as-PCA-CA-PCoA-and-NMDS