Table 3 Non-convergence rate, analysis run time, and number of DEGs for 4 hypothesis tests in the shock dataset. The run time for fitting the full model for each gene, as well as the total time to fit models and perform hypothesis testing is displayed. There were 14,340 genes in the dataset and genes were labelled as a DEG if the Benjamini Hochberg adjusted p-value was < 0.05. For NBMM-LP, the percentage of genes in which one or more reduced models failed to converge is shown in parentheses after the full model non-convergence rate

% Non-converged

0%

0%

0%

0%

0.04%

0%

0.31%

0%

4.33% (9.07%)

0.45%

0%

Analysis Time (minutes)

Model Fit Time

1.14

6.44

0.61

5.8

11.01

2.16

8.91

24.83

326.74

62.72

-

Total Time (Model Fit & Hypothesis Testing)

6.05

28.49

0.72

6.76

18.19

2.2

11.76

104.78

1450.86

65.6

415.12

Between Subject Test

SS vs. CS (Any timepoint)

5,323

-

5,179

-

4,034

3,801

3,009

3,303

3,633

3,351

3,461

Within Subject Tests

Change over time - SS group

5,252

8,331

5,155

8,319

7,910

9,796

8,048

8,600

7,772

8,281

8,081

Change over time - CS group

128

1,381

123

1,313

872

638

1,003

1,836

1,280

1,229

168

Interaction Test

Any Significant Interaction Effect

369

1,842

358

1,783

1,416

1,962

1,502

2,356

1,801

1,769

1,045