Figure 5
From: A Beta-mixture model for dimensionality reduction, sample classification and analysis
Graphical view of the model. The first probe of a probe pair belonging to class (a1, a2) reaches the states L, M and H with probabilities 
, 
and 
, respectively, and each state emits a methylation value from a corresponding beta-distribution (denoted in the figure by 
). Sample index i is suppressed for clarity. Then, transition to the state of the second probe happens according to the transition probabilities 
and similarly to the first probe, a methylation value is emitted from a beta-distribution.