Fig 2. Mutual information performance and set size.

MI and entropy for molecular signatures of increasing size. Points in black represent maximum entropy for a given set size–maximum possible information; points in red are MI values of random samples sets from columns A₁ and A₂ of the GEM (likely to be more similar as they belong to the same cell group); blue measurements are from columns A₁ and B₁ of the GEM (likely to be less similar); Triangle markers represents the molecular signature size minimum used in our calculations; molecular signatures below this size were excluded from consideration. A: zoomed group size 1 to 50; B: a larger plot of group size 1 to 500.