Skip to main content
Fig. 5 | BMC Systems Biology

Fig. 5

From: Uncovering distinct protein-network topologies in heterogeneous cell populations

Fig. 5

Recovering correctly the edge probabilities network between pRaf, ppMek and ppErk for each cell subpopulation in a mixture by UNPBN. The triangles show, color encoded, the posterior edge probabilities between pRaf, ppMek and ppErk at different time points after stimulation as derived by UNPBN analysis of pure NGF-stimulated or EGF-stimulated cell populations (top two rows) with noise level of 0.2 and of a mixture of these populations by GBN versus UNPBN (bottom three rows). Note, for the pure subpopulations, the changes in the edge probabilities between the same components at different time points and upon the different stimulations. Analyzing a mixture of NGF-stimulated or EGF-stimulated cells without unmixing yields uninformative average edge probabilities, not representing any of the distinct two subpopulations. Employing UNPBN recovers precisely the pure edge probabilities for each of the two subpopulations. For clarity, we show here the edge probabilities for 4 time points after stimulation that illustrate the need for unmixing

Back to article page