Fig. 1
From: A mixture model for signature discovery from sparse mutation data
A plate diagram for MMM. The mutations \(o_{t}^{n}\) (1≤t≤Tn;1≤n≤N) in each sample are observed (gray circles) and are modeled as depending on a latent (empty circles) signature ek. The (latent) selected signature \(z_{t}^{n}\) that generated the observed mutation depends on an exposure vector πn that is selected once for each of the N samples