Fig. 4: Illustration of the pipeline employed for case study 2.

We combined prior knowledge TF–TF interactions extracted from METACORE(r) database with variants of Boolean dynamical properties extracted from time series bulk RNA-seq data. We employed BoNesis to sample Boolean networks fulfilling these properties. Then, we performed prediction of reprogramming determinants using CABEAN on a subset of these sampled models. The identified reprogramming determinants are combinations of gene knock-outs and constitute activations. The predictions are then assessed on the other sampled Boolean models using EnsembleMaBoSS simulation, leading to a scoring in terms of fidelity and efficiency, and thus a ranking of the most robust predictions.