Skip to main content
. 2021 Mar 4;12:574529. doi: 10.3389/fneur.2021.574529

Figure 3.

Figure 3

Longitudinal feature-wise analyses of change in MDS-UPDRS Part III in response to Levodopa. MaAsLin2 (Multivariate Association with Linear Models, Version 2: https://huttenhower.sph.harvard.edu/maaslin2), was fit to identify differentially abundant taxa associated with Levodopa use. The MaAsLin2 model was fit with data from all four time points, treating subject as a random effect, and adjusting for age, sex, and BMI. We used the default parameters in MaAsLin2 (https://github.com/biobakery/Maaslin2) .