Amplicon sequencing variant and sequencing depth


I am new to microbiome and trying to understand the basics;

I do have 162 samples and 67405 Number of Sequence variants

and summary(depths) is like below:

Min. 1st Qu. Median Mean 3rd Qu. Max.

18396 77714 103546 112311 135904 308487

I am able to understand it like:

I got total 67405 sequencing variant in 162 samples and each of these sequence variants is unique.

and in feature biome table, I do have the number of each of these(67405) sequence variant how many times detected in each sample, is it right?

depth also represents how deep (how many times), these sequence variant is detected in each sample.

one more query regarding differential abundance:

each sequence variant is unique but more than one can belong to the same species as genus may be different. but when we do differential abundance analysis, do we do it at the level of family or species?

Kind Regards

Good afternoon,


Also right!

Differential abundance is done at the feature level. So for example, you might have two feature that have the exact same taxonomy, including genus and species. However, you know they are not identical because their sequence is different and they are in different features. Maybe they are different strains of the same species. So when you do differential abundance testing, you might find that one strain is higher in one group while the other is the same in both groups.

This is why I recommend doing Differential abundance testing at the feature level.

Let me know if this helps answer your questions,


Thanks Colin, could you please suggest me how we can do differential abundance testing at the feature level, which tools can be used?

I tried to rum ancom command but it never run.
what are the other methods to do this?

Use ANCOM. Input a feature table that has not been collapsed on taxonomy. That's all there is to it.

Unless if you received an error message, chances are it was running... it can take a very long time to run because it is performing a very large number of tests. Filter your feature table to remove low-abundance features before running. You will increase the speed and focus on more abundant features.

If you mean that you received an error message that prevented you from running ancom, that's what we are here for: if you would like help troubleshooting, please report this error message in a new topic.

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