search

FrederickHuangLin / ANCOMBC

Differential abundance (DA) and correlation analyses for microbial absolute abundance data
https://www.nature.com/articles/s41467-020-17041-7
101 stars 24 forks source link

NAs and 1s in ancombc2 primary result table #266

Open lmartinezgili opened 1 month ago

lmartinezgili commented 1 month ago

Dear Huang,

Thank you for the ANCOMBC package.

When I run ancombc2 - with fixed and random effects (in R 4.4.0) - I noticed that for some of the features all the "lfc" columns appear as NA and the "p" and "q_" columns as 1.

In addition, I get a warning message: "In rbind(c((Intercept) = 83.9999999980089, age = 83.9999999980259, : number of columns of result is not a multiple of vector length (arg 18)"

This happened independent of taxa level tested (I tried genus, family and order). I could not find any posted issue describing that, and I cannot see any particular problem with the features that appear as NA. I used only default settings.

Do you know what could be the issue?

Thank you

Maggie8888 commented 1 month ago

It is possible your mix-effect model did not fit your data well. You can try to change the fixed effect variables and random effect variable in your model.