bioinformatics.org/sms/rev_comp.html ]. The pldA alignment was stripped of gaps in BioEdit [51] and imported into MEGA5 [52] for model selection as described above. The alignments were analyzed in PhyML [53] using 1000 bootstraps and the Kimura Selleck RG-7388 two-parameter (K80) model with the gamma distribution (five rate categories) and invariant sites
set to 0.34 and 0.53, respectively; this model was found to be the best by MEGA5. A consensus tree was made in Phylip’s Consense package [54] and represented as an unrooted radial tree in FigTree. The pldA dataset was also analyzed using the same model (GTR + G + I) used for the reference tree. The two pldA trees generated using the GTR + G + I and K80 + G + I models were compared with the TOPD/FMTS software [55]. A random average split distance of 100 trees click here was also created to check if the differences observed were more likely to have been generated by chance. Comparison of pldA sequences with seven core housekeeping genes The average pairwise selleck kinase inhibitor nucleotide identity for pldA and concatenated HK sequences was calculated in BioEdit [51]. The average genetic distance was calculated with the default K80 algorithm in MEGA5 [53, 56]. Horizontal gene transfer analysis of pldA and OMPLA sequences The DNA stability was determined by calculating the GC content of the pldA sequences using SWAAP 1.0.3 [57]. The GC content of
the pldA sequences was compared to the overall GC content of the H. pylori genomes, and significant differences between these two groups
were calculated using a two-tailed t-test (Excel 2003, Microsoft, Redmond, WA, USA). The Codon Adaptation Index (CAI) detects codon bias in a DNA sequence and indicates the possibility of HGT. CAIcal [22] was used to calculate the degree of codon bias and compare it to an estimated value from a reference set ID-8 (eCAI). The OMPLA protein sequences from 171 species were used for an intra-species phylogenetic analysis. Sequences were collected both from the KEGG database [58], using KEGG orthologs belonging to EC13.3.13, and, NCBI’s similar sequence option. Both NCBI Batch Entrez http://www.ncbi.nlm.nih.gov/sites/batchentrez and the Protein Information Resource (PIR) [59] were used to retrieve the protein sequences. Pairwise sequence identities were calculated for ClustalW aligned sequences in BioEdit [51]. Sequences with pairwise identities between 15-90% were kept, and the sequences (Appendix 1 lists all of the Protein IDs used) were re-aligned using the MAFFT web server http://www.genome.jp/tools/mafft/, where the auto-option chose the FFT-NS-i model (an iterative method) [60]. Jalview [61] displayed the minimum, maximum, and average number of residues in the alignment. Poorly-aligned and divergent regions were removed using Gblocks [62].