Table 10.

Summary of the Characteristics of a Parsimonious Phylogenetic Analysis through Polarity Assessment of Gene-Expression Values Followed by a Maximum Parsimony Analysis

• Offers a qualitative assessment of microarray gene expression data by sorting expression values into derived or ancestral.

• Identifies synapomorphies (shared derived expression states) and uses them to delineate clades (class discovery). Synapomorphies are also the potential biomarkers.

• Searches for the most parsimonious classification of specimens; the one with minimum number of steps, reversals, and parallels.

• Efficiently models the heterogeneous expression profiles of the diseased specimens. Those with fast mutation rate such as cancer.

• Incorporates gene expressions that violate normal distribution in a set of specimens—e.g., dichotomously expressed genes.

• Reduces the sensitivity to experimental noise.

• Permits pooling of multiple experiments.

• Allows intra and intercomparability of data.

• Produces higher concordance between gene lists than statistical methods (F & t-statistics and fold change).

• Offers a nonparametric data-based, not specimen-based, gene listing and gene linkage.