In all models, we tested for linear effects, first-order interactions and model fit using the likelihood ratio test.
IC50 values were interpolated from a four parameter logistic model fit using the software EC 50 calculator (see ESM for more details).
Figure 3 is a representative set of time-activity data obtained with a image-derived measured blood input function, which illustrates the good statistical quality of the data and model fit using nonlinear regression and two-tissue-compartment model.
Consequently, we propose to use alternative thresholding strategy based on the model fit using pseudo-False Discovery Rates derived on the basis of the empirical null estimated as part of the mixture distribution.
We further evaluated model fit using weights (WAIC), which are the relative likelihoods of each model given the data.
Based on the marginal structural model fit using targeted maximum likelihood estimation, we estimated the probabilities of virologic failure at each set stratum of adherence, for duration of continuous suppression ranging from 1 to 12 months.
We used the robust multi-array average (RMA) normalization method [36], which consists of three steps: background correction, quantile normalization (each performed at the individual probe level), and robust linear model fit using log-transformed intensities (at the probeset level).
Multi-chip expression intensity normalization was performed for all 4 samples on gene expression chips using the robust multichip average (RMA) algorithm [52], which consisted of three steps: background correction, quantile normalization (each performed at the individual probe level), and robust linear model fit using log-transformed intensities (at the probe set level).
We assessed model fit using the Akaike Information Criterion AICC).
All instruments demonstrated reasonable to good model fit using standard fit indices.
We evaluated the model fit using Akaike's Information Criterion for quasi-Poisson (Q-AIC).
