Fig. 6: GGOT predicts whether an unknown sample reaches the tipping point.

a The unknown sample prediction accuracies of diseases with various progression rates. b The F1 score of different disease prediction results. The results are evaluated from an unbalanced perspective. We further show the predictions of disease stage distribution for a single sample in (c) GSE48452, (d) GSE2565, (e) LUAD, and (f) XJTUSepsis. We denote the tipping point boundary by dashed lines. The left half is the prediction of samples before critical transitions using blue curves and the right half is the predictions after critical transitions using red curves. Different curves indicate the potential distribution of different samples, with stars representing prediction results. GGOT effectively recognizes unknown samples that are approaching the tipping point. The stage distribution of the sample is directly related to the corresponding gene expression.