Figure 5.
Correspondence between the Entropy-SCI-PHY classification of the H559 data set and the robust MCLb-pref partition derived from the in-vitro DNA-binding profiles. Shown is the correspondence in terms of membership in individual subfamilies, for the pair of partitions MCLb-pref (γ = 94, I = 3.2) and the Entropy-H559 classification. The 84 Entropy-H559 subfamilies are listed in rows, in order of decreasing size (1–48 for the 48 nonsingleton subfamilies), with the 36 singleton subfamilies grouped in the bottom row. Individual homeodomains are colored according to their membership in the 55 mouse clusters from the MCLb-pref clustering solution, which contains 33 nonsingleton clusters and 22 singletons. Different colors correspond to different clusters in the MCLb-pref classification. Singleton MCLb-pref clusters are colored in white. Out of the 48 nonsingleton SCI-PHY subfamilies, 35 are either exactly identical to an MCLb-pref cluster (12 subfamilies: #2, 7, 9, 12, 16, 19, 25, 26, 28, 37, 40, 42, 44), or are entirely contained within such cluster (23 subfamilies: #5, 8, 10, 11, 13, 14, 15, 17, 18, 20, 22, 23, 27, 31, 32, 35, 36, 38, 39, 41, 45, 48), whereas the remaining 13 subfamilies group HDs from more than one MCLb-pref clusters (see text).