An Analysis of Variance was run in GenStat [ 42], using the model containing fixed effects for Tissue within Bird and Plate as a Blocking factor.
Specifically, the number of microspheres in each tissue piece was modeled using a Poisson log-linear mixed model containing fixed effects for exposure and random effects for tissue piece and each dog.
Each model contained fixed main effect terms for group assignment (intervention vs. control), time of measurement (treated as a continuous variable with measurement points of baseline, 3, 6, 9, and 15 months), and their interaction.
Models contained fixed effects for treatment arm, calculated adherence in the preceding interval, and their interaction, and random effects for subject.
These models contained fixed effects as described in File S1, but only a single, individual-level, random effect defining individual-level variance (VI, equivalent to individual repeatability) for all traits.
In the model, X is the fixed, known design matrix with 1≤rank(X)≤ q, and B contains fixed, unknown regression coefficients.
The model will contain fixed effects for follow-up time (2 levels) and for intervention group (3 levels), as well as their interaction.
The activity of amlodipine was reproduced in model membranes that contained fixed amounts of charged phospholipid (glycerophospholipid) in a concentration-dependent manner.
Model standard solutions containing fixed amount of 1 mg L−1 La III) with either individual or mixed matrix ions were treated according to the recommended procedure.
