PCoA results interpretation

[Parix]

Hi,

I have performed PCoA analysis using bray-curtis distance matrix and it looks likes this

There is quiet big variance explained by the first axis and there is a clear separation of data along that axis, however, none of the metadata categories can explain this separation. How should one analyse further in this case?

[timanix (Timur Yergaliyev)]

Hi!
What about other beta diversity metrics? Each metric calculated differently from others, and by comparing, which metrics are responding to the treatments, you may gain some interesting insights.

[jwdebelius (Justine Debelius)]

Hi @Parix,

I agree with @timanix, other metrics may give you a different picture. A second question: is this at all related to a technical effect? Are you able to go back and get information about storage, extraction and sequencing? Often a large seperation can be due to technical confounding.

Best,
Justine

[Parix]

Hi,

Thanks for your inputs!

@timanix I see the same trend in weighted unifrac, but not in unweighted unifrac or jaccard distance matrices.