Figure 1.

The multivariate statistical model of the TB, NC, LC, and CAP groups. Scatter plot of OPLS-DA model and validation model of permutation test: (a, b) NC vs. TB: R²X = 0.526, R²Y = 0.983, Q²=0.846, P<0.005. (c, d) LC vs. TB: R²X = 0.371, R²Y = 0.849, Q2 = 0.631, P<0.005. (e, f) CAP vs. TB: R²X = 0.361, R²Y = 0.894, Q²=0.788, P<0.005. During 200 random permutations in each group, P<0.005 showed that no random matrix model had better interpretation rate of the Y matrix than the former OPLS-DA model. The model could effectively distinguish the features of two groups. (A color version of this figure is available in the online journal.)

OPLS-DA: partial least squares discriminant analysis; NC: normal healthy controls; TB: tuberculosis; LC: lung cancer; CAP: community-acquired pneumonia.