Results after taxonomy filtering

User Support
1

Dear qiime2 community,

I’ve come out with a problem I am not able to solve...

I am trying to assess the effect of filtering out all ASVs belonging to a specific Order in my results.
This order is unequally represented between two groups, 27% in group A and 11% in group B (identified as differently abundant with ANCOM).

When the Order is filtered out the alpha diversity within the groups changes differently (diminishes more in the group with a higher relative abundance of this Order). Nevertheless, I don’t see any changes in the beta diversity (Jaccard & Bray-Curtis) comparing before and after filtering results (PERMANOVA and Adonis), although I had to change the sampling depth accordingly to the filtering.

I am wondering if this may be related with normalization, which could be somehow ignoring this Order. However, alpha diversity is affected. I’ve seen in moving pictures tutorial that normalization by rarefaction is totally random.

Any help is very appreciated!

Thanks in advance!

2

Hello @pau,

This makes sense to me!

You don't see big changes, or you don't see any changes? Getting identical results after removing an order is pretty strange, and makes me wonder if that order wasn't really filtered out, or the wrong input file was used, or some technical issue happened. Bray-Curtis is a weighted measure, so removing an order composing 27% and 11% should cause notable changes.

Are you looking at the PERMANOVA and Adonis stat tests only, or also looking at the ordination plots before and after removing this order?

3

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.