Figure 2.

Phylogenetic trees for the alignment regions demarcated by G5 and G7H recombination sites. Non-recombinant, as determined by the manual recombination analysis (see manuscript text), as well as G5 and G7H sequences were included in the phylogenetic tree construction. The designations for the fragments are given at the top, to the left of each tree. Bayesian [35] as well as bootstrapped Neighbour Joining [34], least squares [36], and maximum likelihood [37] trees were constructed for each region. Topologies of the trees generated by the four methods for the same region were same, with exception of how C2 and N sequences related to each other and to L and L-RB isolates from recombination site 'y' to the end of the alignment. Shimodaira-Hasegawa (SH) test [38] was used to select the best of the competing but very similar topologies for each sequence region (date not shown). The tree topology that obtained the highest SH score of 1 is presented for each region. Bootstrap values out of 1000 replicates, produced by the Neighbour Joining and the maximum likelihood methods are given at the nodes, before and after the slanted line, respectively. A star given instead of the number indicates that the respective method did not agree with the topology of the optimal tree identified by the SH test at that particular node. Topologies of all trees were tested against each other and the topologies of the trees presented here were found optimal (SH score: 1). SH scores of 0 were obtained when topologies of the trees from between recombination sites were tested against sequence alignments for the regions on the basis of which the given tree was not generated. The same SH score of 0 was obtained when the tree topologies for the regions from the beginning of the sequence alignment to 'w' and from 'z' to end of the sequence alignment were tested against sequence alignments between the recombination sites. Collectively, these results indicated that the different topologies cannot substitute for each other in explaining the variability of SMV sequences between G5 and G7H recombination sites that we identified.

Gagarinova et al. Virology Journal 2008 5:143   doi:10.1186/1743-422X-5-143
Download authors' original image