Illustrating the function of our proposed model using a synthetic dataset.

A set of expression data for 3 RNA-binding proteins (RBPs) and 5 alternative splicing (AS) events in 100 cancer specimens was simulated. (A) and (E) Original RNA-binding protein (alternative splicing event) expression data. (B) and (F) Simulated cross-sectional RNA-binding protein (alternative splicing event) expression data. The sample IDs of the synthetic data were randomized and the EM transition process information was retained. (C) and (G) Recovered RNA-binding protein (alternative splicing event) expression dynamics according to inferred pseudotime progression trajectory. (D) Comparison of the inferred pseudotime progression with the true progression in the synthetic dataset, evaluated using Spearman’s rank correlation coefficient (rho). (H) Accuracy of the dynamical systems inference evaluated using the AUC of ROC.