Denominator contains only one ASV

I have 16S data from 64 low complexity samples (196 ASVs)that I am trying to analyze with Gneiss. I generated ASV with dada2 and exported the taxa and otu table from R as biom file. I then filtered out ASV with less than 100 reads total, which leaves me with 109 ASVs. I do the pseudocount and build a tree based on hierarchical clustering with the correlation-clustering. When I look at the balances with the balance-taxonomy I consistently see only one ASV in the denominator from y0 to lower levels. Is my data not complex enough for this kind of analysis? Trying to understand why all but one taxa would end up in the nominator for the highest level.

I have only run into this phenomenon, and that happened when we applied hierarchical clustering to the American Gut data. My best guess is that there is a substantial amount of variance between every pair of species.

What does the heatmap look like? That can help with the diagnosis.

HI Jamie,
Here is the heatmap according to a cluster. As you can see the data is very sparse. I tried to also filter out ASV that had a prevalence <5%, but didn't really change much. I think my data is just to variable to really get a signal out of all the noise.

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