Figure 1. Schematic illustration of sample-perturbed Gaussian graphical model (sPGGM) for identifying pre-disease stages. (IMAGE)
Caption
(a) Disease progression can be classified into three states: the normal stage, pre-disease stage and disease stage, with the pre-disease stage representing a critical threshold just before the onset of disease symptoms. (b) The baseline distribution is fitted from reference samples, whereas the perturbed distribution is derived from mixed samples that combine a specific case sample with reference group. (c) The proposed sPGGM constructs candidate detection stages at the single-sample level by utilizing a Gaussian graphical model embedded with prior knowledge of the PPI network and quantifies the distributional changes between the baseline and perturbed distributions through the application of optimal transport theory. Then sPGGM score is used to measure the critical transitions of complex diseases, with a marked increase signalling the pre-disease stage.
Credit
©Science China Press
Usage Restrictions
Use with credit.
License
Original content