Reader Comments

Post a new comment on this article

Tree rooting and recombination

Posted by michaelgaunt on 19 Feb 2014 at 17:37 GMT

The authors believed they had identified robust phylogenetic incongruence between phylogenies. Robust phylogenetic incongruence is strongly compatible with recombination, viz.

"The most striking difference between the S and M segment trees was that phylogroup II (i.e. LQUV and NVAV) were basal in the M segment tree with relatively strong statistical support (Figure 2B) (although it is important to note that we were unable to amplify the M segment sequence from HUPV), while the Soricidae-associated viruses of phylogroup I occupied the basal position in the S segment tree and with much stronger support (Figure 2A). This different phylogenetic pattern was also apparent in the relevant amino acid trees of the N and GPC proteins (Figure S2A–F). Such phylogenetic incongruence is strongly suggestive of reassortment among hantaviruses of phylogroups I and II, and which might have occurred during the evolution of hantaviruses carried by bats and insectivores as these phylogroups are currently only associated with these mammalian species." Guo et al, 2013

The authors need evaluate there assertion against an alternative explanation of a rooting artefact, e.g. mid-point rooting is a common output in tree building programs where an outgroup is not specified. The M tree is rooted with "phylogroup II" and "phylogroup I" is incongruent against the S tree which is rooted with "phylogroup I" and "phylogroup II" is incongruent. If all phylogenies were rooted consistently, e.g. with "phylogroup I" would they then be incongruent? The same argument is true of the amino acid trees in the appendices.

The paper was subject to a student talk at LSHTM's Molecular Virology course, when the matter arose and I promised the class I would raise the issue herein.

No competing interests declared.

RE: Tree rooting and recombination

zhangyongzhen replied to michaelgaunt on 06 Mar 2014 at 02:59 GMT

Dear Dr Michael Gaunt

Thank you for your comments. In our paper, we provided the trees estimated by three methods (BEAST (MCC), Bayesian, and ML methods). Because the MCC tree is automatically rooted on the assumption of a molecular clock, it enables the determination of which viral lineage is most likely to be basal. Accordingly, the basal lineages estimated by the MCC trees were used as an outgroup to root the phylogenetic trees inferred under the Bayesian and ML phylogenetic analyses in our study. In the rooted trees estimated by the BEAST software, phylogroup I and II occupied the basal position in the S and M trees, respectively. In addition, we also used the mid-point rooting when these trees were constructed. The basal lineages estimated by the mid-point rooting were consistent with those in the MCC trees based on the S and M segment sequences. Finally, because only the M segment sequences from Longquan virus are available, further efforts are needed to confirm whether other bat-borne hantaviruses still occupy the basal position in the M tree on a larger sample of bats.

With best wishes,

Yong-Zhen Zhang

No competing interests declared.