search

FrederickHuangLin / ANCOMBC

Differential abundance (DA) and correlation analyses for microbial absolute abundance data
https://www.nature.com/articles/s41467-020-17041-7
104 stars 28 forks source link

Effect size standard error bars span zero #87

Closed eam12 closed 1 year ago

eam12 commented 2 years ago

I suppose this is more of a stats question than anything else, but I have been using ANCOM-BC and discovered that the vast majority of my differentially abundant taxa (q-val < 0.05) have a standard error that spans zero (see image for example). Is this something I should be concerned about or does it just indicate there is a lot of variation in the dataset? Thank you so much! example

FrederickHuangLin commented 2 years ago

Hi @eam12,

Can you check the zero_ind table and see if these differentially abundant taxa (which have standard errors spaning around zero) are considered as structural zeros?

Best, Huang

eam12 commented 2 years ago

Yes, it does appear that the differentially abundant taxa with SE spanning zero are structural zeros. So would the most appropriate thing be to simply remove any differentially abundant taxa identified as structural zeros from further analysis?

FrederickHuangLin commented 1 year ago

Hi @eam12,

I am pleased to share with you that the major update for ANCOMBC package has been completed. Here are some highlighted new features:

We also added the following clarifications for structural zeros:

A taxon is considered to have structural zeros in some (>=1) groups if it is completely (or nearly completely) missing in these groups. For instance, suppose there are three groups: g1, g2, and g3. If the counts of taxon A in g1 are 0, but they are nonzero in g2 and g3, then taxon A will be considered to contain structural zeros in g1. In this example, taxon A is declared to be differentially abundant between g1 and g2, g1 and g3, and consequently, it is globally differentially abundant with respect to this group variable. Such taxa are not further analyzed using ANCOM-BC2, but the results are summarized in the overall summary.

Basically, we recommend separate discussions on taxa with structural zeros (NOT detected by ANCOM-BC or ANCOM-BC2) and taxa without structural zeros (detected by ANCOM-BC or ANCOM-BC2).

I just pushed the changes to the Bioconductor branches. It might take a few business days for the package to become available (but the devel version is ready here!).

Best, Huang

eam12 commented 1 year ago

Wonderful! Thank you so much!