Skip to main content
Fig. 1 | BMC Systems Biology

Fig. 1

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

Fig. 1

UNPBN addresses the challenge of studying intracellular protein networks caused by unmeasured proteins and inter-cellular heterogeneity. a A biochemical system for which three proteins (x, y, z) are being measured in the same cell while the other proteins are unmeasured. Note that the effects of z on x are mediated by unmeasured proteins (α and β). b Depending on the level and state of these unmeasured proteins, the measured causality between x and z can differ qualitatively between cells. For example, normal and cancer cells have different activity levels of oncogenes and tumor suppressors which here lead to a negative or a positive causal effect of z on x, respectively, thereby to a controlled growth or a constitutive growth. c Left, multiparametric high-throughput single-cell measurements (e.g., flow-cytometry) of a heterogenous sample of cells (e.g., cancer and normal cells). Middle, attempts to statistically infer a single set of relations (here, causal topology) between the measured proteins fail because there are two distinct subpopulations having two distinct sets of relations. At the same time, it is also impossible to identify the two distinct subpopulations as two distinct proximity-based clusters. Right, UNPBN performs unmixing and inference of statistical relations as one process, thus finds the set of sets-of-relations (network topologies) that explains best the observations

Back to article page