fudan.edu.cn/cvtree/) with the K parameter selleck kinase inhibitor set at 6 [17]. The outcome from the program is a distance matrix based on amino acid sequence comparisons, which is then used to generate a phylogenetic tree with the neighbor-joining method. In the shown tree, the outgroup chosen was Methanothermus fervidus (an Archaea). After tree visualization with MEGA5, branches were collapsed wherever possible with the exception of the Negativicutes branch, which remained expanded. Consensus tree of conserved genes Using the list of universally conserved core genes, previously identified by Ciccarelli et al. [18], and an implementation of BLAST, a set of genes that was shared among all 145 genomes was identified. Proteins that had no match in at least one genome or showed poor E-value were eliminated.

The 27 conserved core genes were extracted (Table 1) and a multiple alignment was produced using MUSCLE software [19]. A set of phylogenetic trees was constructed by PAUP [20] and a best-fit consensus tree was generated using Phylogeny Inference package (PHYLIP) as described elsewhere [21]. Bootstrap values were found after 27 re-samplings, which is equal to the number of gene families conserved in all the analyzed genomes. DNA tetramer analysis and amino acid usage A tetramer frequency heatmap was constructed from the observed ratios of tetra-nucleotide frequencies divided by estimated tetra-nucleotide frequencies for each genome [22]. The estimated tetra-nucleotides were computed from the genomes’ base composition.

The ratio of observed over expected frequency was used for hierarchical clustering using complete linkage and Euclidean distance, which was subsequently performed with respect to both strain and tetramer frequencies. The amino acid heatmap is based on frequencies of deduced proteomic amino acids from each genome normalized with respect to the total number of amino acids in each genome. The amino acid frequencies for each genome were clustered using complete linkage and Euclidean distance with respect to both genomes and amino acids. The heatmap was made using the R package ggplot2 [23]. Comparison of metabolism potential The protein sequences of Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology categories [24] were downloaded and only the Bacterial sequences were considered.

The Hidden Markov model (HMM) of each ortholog was generated using HMMER version 3 [25] based on the multiple alignment of each orthologous set of KEGG proteins, using MUSCLE software [19]. The 145 proteomes were queried against the HMMs to infer their ontology. A cutoff GSK-3 of 1��10?30 was used for statistical significance. A heatmap of each pathway and process derived from the database KEGG was illustrated based on normalized abundance of the enzymes present in each pathway.