Is there a simple way to convert between similarity and disimilarity

[Tina_Khone]

Hello!

Is there a way to check for similarities rather than differences between groups in Qiime2?

Thanks

[jwdebelius (Justine Debelius)]

Hi @Tina_Khone,

This obviously depends on your metric. For metrics between 0 and 1, you can just subtract the metric from 1, so

similarity = 1 - dissimilarity, for 0 ≤ dissimilarity ≤ 1.

I'm not sure what to recommend for metrics/indices which are not bounded.

Best,
Justine

PS I also edited your title to make it easier for other people to find, I hope that's okay!

[Tina_Khone]

So if I want to check which bacteria are similar between the groups, which data should I use? the taxa bar plot?

[jwdebelius (Justine Debelius)]

Hi @Tina_Khone,

This is a really different question. It depends on how you want to define "similarity" and at what level. I'm not a big fan of feature bar plots - they're a great diagnostic but often offsicate the low level changes where things happen. You could maybe look at the idea of the "core" microbiome, keeping in mind that "core" is going to vary from system to system and taxonomic level to taxonomic level.

Best,
Justine