Metaves/Coronaves (1 Viewer)

[l_raty]

Would love to get thoughts about the attached tree...
This tree computed from sequences of introns 4 and 5 of the Fibrinogen Beta Chain (FGB/b-Fib), instead of the more widely used intron 7, from which the Metaves/Coronaves hypothesis was originally derived (data mainly from Hackett et al's 2008 Science paper, with a couple of additional taxa from GenBank; the analysis is ML, model TVM+G selected based on the AICc criterion, supports based on 100 bootstrap replicates; both introns gave congruent results when analyzed separately, thus I simply concatenated them [without the intervening exon]).
Besides the general subdivisions in Neoaves, note the positions of the two tropicbirds and of the hoatzin.

 

[Peter Kovalik]

Very interesting.

 

[l_raty]

I'm more interested in the broad implications.

Even if we discard the fact that half of the Coronaves clade identified by intron 7 appears closer to the bulk of Metaves based on introns 4-5 (BS there is 'only' 84%), intron 7 places Phaethon and Opisthocomus quite unambiguously in the Metaves radiation, while introns 4-5 place them in the other half of Coronaves with a 100% BS. The only straightforward way I see to reconcile these two results is genetic recombination - at the onset of the radiations of Metaves and of this subclade of Coronaves, birds must have existed that carried alleles of both lineages (so that a crossing-over in the gene, during a meiosis in a heterozygous individual, could result in the creation of a new, chimeric allele), either as a result of a long history of genetic polymorphism, or as a result of interbreeding barriers between populations carrying each lineage having broken down.
What this would mean, is that the entire Metaves/Coronaves hypothesis rests on a gene giving a signal that, at this level, almost certainly departs (to an extent that is probably hardly assessable) from the taxon tree...

Or is there another possible interpretation?

Laurent -

 

[Mysticete]

the key thing here than is that obviously we need to sequence more nuclear DNA for birds, and see how well the Metaves and Coronaves hold up...

 

[l_raty]

the key thing here than is that obviously we need to sequence more nuclear DNA for birds, and see how well the Metaves and Coronaves hold up...

With the caveat that, if really what happened at the base of Metaves was a rapid radiation, the real phylogenetic signal might be extremely weak. In a combined analysis, even with a lot of DNA, a very weak phylogenetic signal is unlikely to override a strong, but non-phylogenetic signal present in one of the genes...

 

[Mysticete]

however the other genes may not possess such a strong non-phylogenetic signal. Certainly increasing gene sampling has helped resolve some of the issues with whales

 

[l_raty]

I'm not saying that increasing sampling is not a good thing.
What I'm saying is that the expected 'signal' for a rapid radiation is very little or no signal at all (which is basically what has been found up to now at the base of Neoaves in all genes but FGB). If you have a lot of data with no/very little signal, then add to these a single gene that has a strong, but non-phylogenetic signal due to a violation of orthology, the non-phylogenetic signal is not going to be easily buffered by what is in the remaining data. Thus the fact that this signal holds up in a phylogenomic-type analysis is not necessarily a good indication that it is correct.
It has been argued that genes likely to exhibit a non-phylogenetic signal should be discarded from the analyses (e.g., Jeffroy et al. 2006).

 

[mb1848]

The dynamic evolutionary history of the bananaquit (Coereba flaveola) in the Caribbean revealed by a multigene analysis used the fifth intron Fibrinogen
http://www.biomedcentral.com/1471-2148/8/240 .
Looked interesting:
A molecular phylogenetic survey of caprimulgiform nightbirds illustrates the utility of non-coding sequences 2009
http://www.sciencedirect.com/scienc...serid=10&md5=ad1fbc31988ff2235fc8ba998ba65443