error in linear-mixed-effects: 'singular matrix' in nested design


I am trying to run longitudinal linear-mixed-effects, however, due to my nested design, I am getting the error: "Linear model will not compute due to singular matrix error."

I know my "Donation" is dependent of the "Individual" variable, however, I am not sure how to proceed (Donation is a nested factor of Individual). My goal is to be able to generate a volatility plot similar to this (, where I can plot the change in distance over time, per "Individual".

Here is my code:
qiime longitudinal linear-mixed-effects
--m-metadata-file bray_tp_baseline.qza #output of first-distances
--m-metadata-file meta_long2.tsv
--p-metric Distance #bray
--p-state-column TP #Timepoint 1, 2 or 3 (some samples have 1&2 timepoints, some have 1,2&3 timepoints)
--p-individual-id-column Donation #some individual produce multiple donations, nested factor of Individual
--p-group-columns Individual #different individuals
--o-visualization bray_tp_lme.qzv

metadata :
meta_long2.tsv (929 Bytes)

Any help will be greatly appreciated :slight_smile:

Thank you!

Looks like you should switch Individual and Donation factors since Individual is a column with ID and Donation is a group. Not sure if it will work as you intended but give it a try.


1 Like