Figure 1
From: A human genome-wide library of local phylogeny predictions for whole-genome inference problems
Illustration of phylogenetic imperfection. A hypothetical maximum parsimony (MP) tree created from the input sequences 000, 110, 101, and 111, which can be considered binary representations of the alleles at three SNP loci in four individuals. Each edge is labeled with the SNP locus that mutates along that edge. The tree has imperfection 1 because the number of mutations required to explain the tree (4) is one more than the number of variant sites (3), manifesting as two variations in site 3 (dashed circles). Note that imperfection is distinct from the number of unobserved (Steiner) nodes that must be inferred to fit the tree; one such node (100, in grey) is required for this tree. Note also that the MP tree need not be unique for a given data set, but the imperfection score will be the same for all MP trees.