Computational protocol: Evolutionary history of rat-borne Bartonella: the importance of commensal rats in the dissemination of bacterial infections globally

Similar protocols

Protocol publication

[…] The citrate synthase gene, gltA, is an especially popular and widely used molecular target for distinguishing between closely related Bartonella species and genotypes. Most laboratories working with Bartonella bacteria have successfully used this genetic marker, and consequently the majority of Bartonella sequences submitted to GenBank are derived from a 338 bp fragment of this gene. In addition to using sequences from GenBank, we were able to analyze sequences from the large database of the CDC Bartonella Laboratory (Fort Collins, CO), many of which have not been published. Therefore, we used 131 gltA sequences of B. elizabethae-CSL () from 15 Rattus or Bandicota rat species (“true rats”) from 17 countries (), and 191 gltA sequences, including the previous 131 from true rats, of B. elizabethae-CSL () from 27 rodent species from the Murinae subfamily from 17 countries in our analyses.Sequences were aligned using ClustalX2.1 (Larkin et al. ). Sequence analysis was undertaken using Bayesian Markov chain Monte Carlo (MCMC) phylogenetic analysis using BEAST software (version 1.6.1) (Drummond and Rambaut ). Due to the nucleotide sequences being from coding regions, we used the SRD06 model; (Shapiro et al. ). The SRD06 model uses the Hasegawa, Kishino, and Tano (HKY) model of substitution with gamma (Γ)-distributed variation in rates among sites. Base frequencies were estimated and because of the low diversity in the gltA gene we estimated the nodes of the tree using substitutions per site. We assumed a constant population size and used uninformative priors throughout the analyses. A 120 × 106 MCMC chain length was selected in BEAUti (version 1.6.1) and run in BEAST (version 1.6.1). Parameters were logged every 12000 trees. Models were accepted if parameter effective sample sizes (ESS) were >200 (but see Results and Discussion) and MCMC chains for parameter estimates had converged (Tracer, version 1.5). As per standard practice in MCMC phylogenetic analysis, the first 10% of maximum clade credibility (MCC) trees were discarded in TreeAnnotator (version 1.6.1) as burn-in before the chain has converged. Resulting MCC trees were visualized in FigTree (version 1.3.1). Sequence traits were categorized by state variables, and we used subregions within Eurasia (Southeast Asia, East Asia, South Asia, Western Asia, and Western Europe) and Africa (East, Central, and West), but did not subcategorize Australia and North or South America. We tested the hypothesis that these bacteria are derived from Asia and dispersed globally from there. We performed this analysis on two data sets, the first including B. elizabethae-CSL sequences derived only from true rats (Rattus and Bandicota genera), and the second from the wider Murinae subfamily. We also categorized the sequence traits based on the host from which they had been isolated (), in order to test the hypothesis that commensal rats, R. rattus and R. norvegicus, were involved in the ongoing transmission of B. elizabethae-CSL throughout the globe. […]

Pipeline specifications

Software tools Clustal W, BEAST, FigTree
Application Phylogenetics
Organisms Rattus norvegicus, Homo sapiens, Rattus rattus
Diseases Bacterial Infections, Bartonella Infections