In phylogenetics, the statistical principle used to create the phylogenetic tree (phylogram) most likely to account for the observed data. Methodologies based on this principle are frequently applied in computer analysis of molecular sequence data and require a mathematical model describing how DNA or protein sequences behave. For example, the model will account for how bases change at different rates depending on the particular gene, the position in a codon, and whether the change is a transition (i.e. between two purines or two pyrimidines) or a transversion (i.e. purine to pyrimidine or vice versa). The assumption is that evolutionary histories of organisms adhere to such rules.