Cladistic coding of genomic maps

摘要

A new method of genomic maps analysis is described. The purpose of the method is to reconstruct phylogenetic relationships from the geomic organization of taxa. Our approach is based on gene order coding. This coding allows the description of genome topology without a prior hypothesis about evolutionary events and phylogenetic relationships. Different characters are used for each gene: (1) presence/absence, (2) orientation, and (3) relative position. The relative position of a particular gene inside the genome is the pair of genes surrounding it. The relative position character represents all the positions of a gene in the sampled genomes. It is coded as a multistate character. Our coding method has a priori variable cost implications on operators such as inversion, transposition, and gene loss/gain, which we discuss. The overall approach best fits the "duplication, random loss" evolutionary model. The coding method allows the reconstitution of a possible hypothetical common ancestor genome at each node of the tree. This reconstitution is based on the character states' optimization; it comes down to choosing, among all possible optimizations, the optimization compatible with a complete genome topology at each internal node. The multistate coding of gene relative position, which is an undeniable advantage of this method, permits this reconstitution.