[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Rép : Post–K-Pg Radiation of Placentals -- open access!!!



Tim Williams <tijawi@gmail.com> wrote:

> Yes; although apparently this topology is sensitive to the analysis
> used.  Bayesian recovers these two (_Icaronycteris_ and
> _Onychonycteris_) as stem-chiropterans (stem-bats).
>
> A similar thing happens to _Archaeopteryx_ vis-a-vis avialans (birds)
> when Bayesian analysis is used, IIRC (from one study, anyway).

AFAIK there have been only three Bayesian phylogenetic analyses of
non-neornithine coelurosaurs (Snively et al. 2004; Clarke & Middleton
2008; Lee & Worthy 2011), and all of them found _Archaeopteryx_ in its
usual position (as a sister group to all remaining avialans) despite
some strange results in other parts of the tree. One of these analyses
(Lee & Worthy 2011) supported a topology that was, if anything, even
more pectinate than usual, with _Wellnhoferia_ more closely related to
avebrevicaudans than to _Archaeopteryx_. Something analogous to the
situation with _Icaronycteris_ and _Onychonycteris_ (a clade
comprising _Archaeopteryx_ and _Jeholornis_) was recovered by
Prieto-Márquez et al. (2011), but that was a parsimony analysis.

David Marjanović <david.marjanovic@gmx.at> wrote:

> That may be due to the known tendency of Bayesian inference, when given a
> flat prior on tree shape, to gather basal twigs into larger clades to make
> the tree more symmetric.

I don't think that's correct -- flat priors are "flat" precisely
because they assign the same probability to every possible topology
(tree shape), be it symmetric or not. If a prior biases the analysis
in favor of more symmetric trees, it's no longer flat. It's true,
though, that priors that place equal probability on all _labeled
histories_ (= topologies with internal nodes ordered in time), such as
the Yule process, give more probability to symmetric trees than to
asymmetric ones -- that's because the former are compatible with more
labeled histories than the latter (Velasco 2008). But that's a
property of a particular prior probability distribution, not of
Bayesian inference as such. Unfortunately, O'Leary et al. don't
mention their choice of priors in the supplementary information, but
with such a large amount of data it may be irrelevant anyway.


Refs:

Clarke JA, Middleton KM 2008 Mosaicism, modules, and the evolution of
birds: results from a Bayesian approach to the study of morphological
evolution using discrete character data. Syst Biol 57(2): 185-202

Lee MSY, Worthy TH 2011 Likelihood reinstates _Archaeopteryx_ as a
primitive bird. Biol Lett 8(2): 299-303

Prieto-Márquez A, Minjin B, Horner JR 2011 A diminutive deinonychosaur
(Dinosauria: Theropoda) from the Early Cretaceous of Öösh
(Övörkhangai, Mongolia). Alcheringa 36(1): 117-36

Snively E, Russell AP, Powell GL 2004 Evolutionary morphology of the
coelurosaurian arctometatarsus: descriptive, morphometric and
phylogenetic approaches. Zool J Linn Soc 142(4): 525-53

Velasco JD 2008 The prior probabilities of phylogenetic trees. Biol
Philos 23: 455-73

-- 
David Černý