Skip to main content


Figure 2 | Genome Medicine

Figure 2

From: Computational purification of individual tumor gene expression profiles leads to significant improvements in prognostic prediction

Figure 2

Graphical model representation of the ISOpure model. This directed graph represents the joint probability distribution over the input data (blue circles) and the estimated model parameters (white circles), conditioned on the estimated model hyper-parameters (white circles), in the ISOpure model (equations 3 to 9). The conditional probability of each variable depends only on its parents in the graph. To avoid explicitly representing all N tumor profiles c n and R normal profiles b r in the graph (and other associated variables), we used 'plate' notation by drawing one representative node per variable, and enclosing these variables in a plate (rectangular box), where the total number of instances of each enclosed variable is indicated by the fixed constant in the upper right corner of the plate.

Back to article page