To check for the robustness of this statistical model, we repeat the regression with three of the four dependent variables interacting with the indicator variable for large general merchandise workers.
To test whether the differences in the estimated distortion matrices are caused by the differences in the conductivity models, we repeat our above simulations once again in a homogeneous half-space model with conductivity of 5.13×10−4 S/m.
The second box plot is for 42 choose 2, since there are 861 different combinations, we shuffle the list of all the 42 probesets, then we choose two probesets at random and run the model, we repeat this process 42 times, thus each box plot in Figure 4 has 42 data points.
To account for uncertainties in our model, we repeat our experiments under varying assumptions about (parameter values of) the underlying population genetics and mechanistic models.
To reliably validate our detection model, we repeated all our experiments with 20 different random seeds.
To test whether we were over- or under-fitting our model, we repeated the optimization by averaging the top 20 and 40 genes instead of 30.
To determine the impact of the genetic data on the model, we repeated the analysis using only the "conventional" variables (age, duration of diabetes, diabetes type, sex and ethnicity).
First, to further confirm tear secretion is chronically reduced in this model, we repeated treatments corresponding to the work-rest-sleep cycle of office workers on a daily basis up to 20 days (Figure 1B).
In order to enhance the face validity of our model, we repeated analyses using standard linear regression with baseline adjustment.
To determine growth-limiting nutrients for each model, we repeated the analysis with iSS1393 and iAH991 individually (Fig. 1C).
