Re: Molecular taxonomy for dummies?

[Roberto Takata <rmtakata@gmail.com>]

http://bibiserv.techfak.uni-bielefeld.de/gcb04/tutorials/hoef-emden/tutorial-abstract.pdf
----------

[]s,

Roberto Takata

On Mon, Sep 26, 2011 at 9:03 AM, Dr Ronald Orenstein
<ron.orenstein@rogers.com> wrote:
> In the course of following (at some considerable distance) the discussion on 
> Passeriformes, I am increasingly realizing that I need a basic, simple guide 
> to the subject of molecular taxonomy. I mean something directed at a poor 
> ignorant fellow like myself who hasn't the faintest idea what a retroposon 
> is, or who thinks that long branch attraction has something to do with 
> romance. Does such a thing exist?
>
> Ronald Orenstein
> 1825 Shady Creek Court
> Mississauga, ON
> Canada L5L 3W2
>
> On 2011-09-25, at 4:19 PM, evelyn sobielski <koreke77@yahoo.de> wrote:
>
>>
>>>> Although the taxon sampling is a bit patchy (e.g., no
>>> representatives of
>>>> Acciptriformes and Strigiformes)
>>
>> You need at least 2 of _Sagittarius_, _Pandion_ and elanids plus 2 of 
>> eagle/hawk/buteo. One of _Sagittarius_/_Pandion_ will usually go rogue.
>>
>>> Isn't that rather strange if one is interested in the
>>> position of falcons?
>>
>> Coliiformes and _Aegotheles_ (haven't seen _Steatornis_) give the strongest 
>> bogus signal. Coliiformes are hopeless via DNA. But the framework now exists 
>> for morph. _Aegotheles_ is... funny. I wouldn't be surprised if _Steatornis_ 
>> was no better. Something the entire base of the cypselomorphs has a high 
>> probability to be rogue (Apodiforms are the only thing I could get to clade 
>> dependably)
>>
>> Seriemas pose much the same problem as mousebirds. Their relationship based 
>> on DNA is simply not to be taken seriously. They were one of the *major* Pg 
>> radiations, so this needs to be reanalyzed. You could leave out fossils *if* 
>> they behaved predictably on DNA, but they don't. (Any more seriema DNA is 
>> gladly appreciated though) They are not major upsetters though.
>>
>> Psittaciforms, accipitrids, falconids and passerines are in the same major 
>> clade it looks. Strigiforms probably too. Other than these, "picocoracines". 
>> It looks like 2 subsequent radiations, one is the basal split of Neoavesa 
>> into - basically - Aequornithes and "higher landbirds" and *perhaps* 2-3 
>> more distinct lineages surviving from that time.  Some taxa are "weird" 
>> though - all "Metaves", essentially, though *most* do in fact reliably go in 
>> one clade or the other for *most* loci. The second radiation is very 
>> explosive and somewhat obfuscated (but not usually destabilized) in 
>> Aequornithes by ciconiids etc, in the other clade by upupiforms.
>> As I said, mousebirds and _Aegotheles_ usually manage to completely upset at 
>> least one of the 2 major clades by their sheer presence.
>> I woudn't trust anything that doesn't compare the tree with and without them.
>>
>> Falconids tend to turn up close to passerines, and perhaps reliably so. 
>> Psittaciforms are perhaps more often among Aequornithes than close to 
>> falconids, and if not rogue they are usually closer to accipitrids.
>>
>> NWvultures clade irregularly, with about even likelihood for them to turn up 
>> next to falconids, accipitrids or within Aequornithes. Removing them, 
>> however, usually has little effect on the tree.
>>
>> Both radiations must have occurred in the late K. We have some fossils which 
>> very much look as if they'd turn out on one "side" of charadriiforms, 
>> Aequornithes etc. or the "opposite" one, so the basal radiation of these was 
>> very likely already ongoing.
>>
>> There are regions all over the genome where oligo-indels "happen", i.e. they 
>> are not repaired (conspicuous in an alignment because there are many 
>> oligo-gaps and consensus is low).
>> There are other regions where there is a higher probability of long indels 
>> (usually ins rather than dels). They are rarely phylogenetically informative 
>> per se. A certain insert (and perhaps deletion) may consist of several 
>> independent elements; I found one of the retroposons apparently a fusion of 
>> 4 elements, 2 of which were ubiquitious in Neornithes, 1 has human(?) and 1 
>> was likely protist.
>>
>> In strigiforms, it *may* be necessary to always sample _Phodilus_.
>>
>> I have not found a viable solution for cypselomorphs yet. This and the low 
>> taxon sampling among "picocoracines" is the present obstacle on resolving 
>> the "higher landbirds". And of course, either the tree with a mousebird or 
>> without one (two doesn't help either, not even when both genera are sampled) 
>> or neither may be true. We cannot tell.
>>
>> The two most common taxa to stand basal in Neoaves (mis-rooting to 
>> Galloanseres/paleognaths) are passeriforms and columbiforms. I haven't 
>> really checked what the latter are closest to when you unroot Neoaves 
>> (passeriforms are usually close to something "higher landbird", not rarely 
>> to falcons), but IIRC they *might* be Aequornithes. (Need to add 
>> charadriiiforms and some more Aequornithes to the sample now.
>>
>> 1. Has anyone ever tested the population genetics of 
>> retroelements/transposons? (Should be possible in chicken)
>> 2. What were the BLAST settings used in Suh et al for checking the 
>> retroposon sequences against nr/nt?
>> 3. What specimens were used from _Falco_? Wild or captive? If captive, how 
>> many generations from the wild? (the human-like piece is... odd)
>> 4. What does a cladistic analysis of the more widespread transposons' 
>> sequence say? (Some of the more ubiquitious long transposons seem to carry 
>> significant phylogenetic signal)
>>
>> There is something odd in the pattern of rogue taxa. Can't lay my finger on 
>> it, but in some cases I could fix it by sampling a close relative instead. 
>> In these cases there were often massive indels in the rogue taxa (in one 
>> trochiline genus there was 1 rogue and 1 unrogue sp.). Not finding them in 
>> close relatives would mean that such indels, which are assumed to persist in 
>> situ through evolutionary time with high probability, don't.
>>
>> -----
>>
>> PNAS has a new paper about avian extinction at the K-Pg boundary. Looks 
>> interesting in the news, anyone seen it yet?
>>
>>
>>
>> Regards,
>>
>> Eike
>