Escherichia coli molecular phylogeny using the incongruence length difference test

Molecular phylogeny of the species Escherichia coli using the E. coli refer ence (ECOR) collection strains has been hampered by (1) the absence of root ing in the commonly used phenogram obtained from multilocus enzyme electrop horesis (MLEE) data and (2) the existence of recombination events between s trains that scramble phylogenetic trees reconstructed from the nucleotide s equences of genes. We attempted to determine the phylogeny for E. coli base d on the ECOR strain data by extracting from GenBank the nucleotide sequenc es of 11 chromosomal structural and 2 plasmid genes for which the Salmonell a enterica homologous gene sequences were available. For each of the 13 DNA data sets studied, incongruence with a nonnucleotide whole-genome data set including MLEE, random amplified polymorphic DNA, and rrn restriction frag ment length polymorphism data was measured using the incongruence length di fference (ILD) test of Farris et al. As previously reported, the incongruen ce observed between the gnd and plasmid gene data and the whole-genome data was multiple, indicating numerous horizontal transfer and/or recombination events. In five cases, the incongruence detected by the ILD test was punct ual, and the donor group was identified. Congruence was not rejected for th e remaining data sets. The strains responsible for incongruences with the w hole-genome data set were removed, leading to a "prior-agreement" approach, i.e., the determination of a phylogeny for E. coli based on several genes, excluding (1) the genes with multiple incongruences with the whole genome data, (2) the strains responsible for punctual incongruences, and (3) the g enes incongruent with each other The obtained phylogeny shows that the most basal group of E. coli strains is the B2 group rather than the A group, as generally thought. The D group then emerges as the sister group of the res t. Finally, the A and B1 groups are sister groups. Interestingly, the most primitive taxon within E. coli in terms of branching pattern, i.e., the B2 group, includes highly virulent extraintestinal strains with derived charac ters (extraintestinal virulence determinants) occurring on its own branch.