Interpretation of gneiss lme-regression output

[cdl]

Hi,

I am quite new working with lme-regression in gneiss. I have data from the same individuals taken at 7 different time points (so I chose lme-regression instead of ols-regression). I used performed gradient clustering based on the day the samples were taken (Time_days) and used --p-formula Time_days and --p-groups ‘Subject_ID’ in lme-regression.

I got a FDR corrected coefficient p-value of 2.52EXP-12 for balance ‘y1’ at the species level with 20 taxa in the numerator and 122 in the denominator.

What does exactly the coefficient for that balance ‘y1’ mean?
Group variance = 1.275354
Intercept = 0.370015
Time_days = 0.027925

Also, is there any way I can see using gneiss the variance explained (R2) by that particular balance?

Thank you a lot in advance,
Celia

[mortonjt (Jamie Morton)]

Hi @cdl, note that we are in the middle deprecating LME regression in favor of other differential abundances methods such as songbird and aldex2. See this post as well: Linear Regression Summary in Gneiss

Concerning your other questions, y1 is the name of the log ratio. See the original gneiss tutorial on this https://docs.qiime2.org/2019.10/tutorials/gneiss/

The coefficients are slopes (i.e. differences across groups) in units of log-fold change (also covered in the tutorial).

It’s not clear how to compute R2 in LME, this is an open statistics question as far as I am aware.

[cdl]

Hi @mortonjt,

Thanks a lot for the clarifications. I will also have a deep look to the other differential abundance methods you proposed.

Best,
Celia