A phylogenomic approach to bacterial phylogeny: evidence of a core of genes sharing a common history

Abstract

It has been claimed that complete genome sequences would clarify phylogenetic relationships between organisms, but up to now, no satisfying approach has been proposed to use efficiently these data. For instance, if the coding of presence or absence of genes in complete genomes gives interesting results, it does not take into account the phylogenetic information contained in sequences and ignores hidden paralogies by using a BLAST reciprocal best hit definition of orthology. In addition, concatenation of sequences of different genes as well as building of consensus trees only consider the few genes that are shared among all organisms. Here we present an attempt to use a supertree method to build the phylogenetic tree of 45 organisms, with special focus on bacterial phylogeny. This led us to perform a phylogenetic study of congruence of tree topologies, which allows the identification of a core of genes supporting similar species phylogeny. We then used this core of genes to infer a tree. This phylogeny presents several differences with the rRNA phylogeny, notably for the position of hyperthermophilic bacteria.

Figure 1

Construction of supertrees by MRP with bootstrap weighting. Each tree obtained for a set of species from a single orthologous gene family was coded into a binary matrix of informative sites. Only branches having a bootstrap value (or RELL-BP value for ML trees) over 50% were coded. The matrices obtained were concatenated into a supermatrix in which species absent from a gene family are encoded as unknown state (“?”). The supertree was calculated on the supermatrix using DNAPARS with all default options, and 500 replicates of bootstrap were made using SEQBOOT.

Figure 2

Supertrees of 45 species constructed with 730 trees. (A) Supertree based on trees made by BIONJ and a gamma distribution estimation of evolutionary rate heterogeneity. (B) Supertree made with ML trees. Only bootstrap values over 50% are shown.

Figure 2

Supertrees of 45 species constructed with 730 trees. (A) Supertree based on trees made by BIONJ and a gamma distribution estimation of evolutionary rate heterogeneity. (B) Supertree made with ML trees. Only bootstrap values over 50% are shown.

Figure 3

Plot of the two first axes of the PCO made from 310 BIONJ trees compared with RF distance. The trees chosen contained at least 10 bacterial species. The same experiment with ML-trees gave very similar results. Black dots correspond to informational genes, and gray dots correspond to operational genes. The ellipse contains the 121 trees retained for supertree reconstruction (see Table 1).

Figure 4

Supertrees of 45 species built with the trees selected using the PCO results. (A) Supertree based on 121 trees made by BIONJ and a gamma distribution estimation of evolutionary rate heterogeneity. (B) Supertree based on 118 ML trees.

Figure 4

Supertrees of 45 species built with the trees selected using the PCO results. (A) Supertree based on 121 trees made by BIONJ and a gamma distribution estimation of evolutionary rate heterogeneity. (B) Supertree based on 118 ML trees.