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

Re: Shaking up the bird family tree



> --+--A 1
>   `--+--B 0
>      `--+--C 0
>         `--+--D 0
>            `--+--E 1
>               `--F 1
> 
> Here, state 1 is an unambiguous synapomorphy of E and F.
> That it also 
> appears in A does not matter. "Unambiguous" does
> not mean "unique".

Ah, I meant "unique".

> Isn't 21 bp way too short for a transposon? Surely
> transposase has more 
> than, like, five amino acids?!? And why should a transposon
> insert into the 
> exact same place again and again and again?

4 is really short indeed. 21 is so-so. Very short translocations are rare in 
animals but IIRC quite common in plants. If it were important, I'd check 
whether the 21bp sequence occurs anywhere else in avian genomes, and where. In 
any case, 21bp is too long to be a simple replication error. Even 4bp is 
probably too long to be a simple replication error. Perhaps recombination 
errors, the actual sequence would be helpful.

It's not just the insertion occurring over and over again. The key source of 
error might actually be the redeletion. If you compare close relatives (e.g. 
chimps and humans) you'll find lots and lots of shared insertions. If you 
compare distant relatives (say horses and mice), you find the number much 
reduced.

So some loci are more likely than others to get a (longer, i.e. non-replication 
error) insertion, and some of these are FAR more likely than others to keep it 
for long.

The number of possible insertion sites (transopons or not) is huge, but the 
number where insertions stay *for long* (say 10-20 Ma+) is apparently pretty 
restricted. That much seems to be clear. But functional genetics is required to 
find out why I guess... it is logical to assume that at most sites insertions 
carry a fitness cost that in the long run but not immediately is prohibitive, 
and so they are liable to be lost again.

I would simply not place much weight on transposon/insertion studies until it 
is clear what guides the evolutionary behavior of (especially) insertions 
(transposons or otherwise). Essentially, presuming a constant gap cost is like 
presuming a transition:transversion ratio of 2:3. But it cannot be sorted out 
that easily, except for point insertions (replication errors) where it is 
reasonably close to the truth.

One might check out LOTS of orthologs for insertions, and analyze the inserted 
sequences themselves.

> > Changing the alignment algorithm (in BLAST terms:
> using PAM and not 
> > BLOSUM) would probably prevent you from ever finding
> this supposed 
> > dichotomy. Alternatively, one might treat indels as a
> single character 
> > each - which they technically *are*. In that case too,
> the signal would be 
> > easily overwhelmed.
> 
> Someone should test that (by setting the gap extension cost
> to 0).

It's the variable gap-cost *and* two different approaches. One that searches 
for overall similarity and one that starts with an aligning site and searches 
around it, trying to discover continuous stretches of high sequence similarity. 
The gap penalty is applied in either case. When you do for example functional 
gene analysis, the former is fairly useless; it sucks at finding shared motifs 
and at least across bioinformatics as a whole it is not used very much these 
days.

See also http://larsjuhljensen.wordpress.com/tag/phylogeny/ 

For deep-time analyses of sequence data, it might be better to use point rather 
than block substitution-finding algorithms, because the signal is VERY noisy 
and any indication of relationships is not likely to be conserved as 
microcontinents of alignment in an ocean of nonalignment, but more like a few 
small islets of alignment here and there. But I don't think this has ever been 
tested. I might ask at the chromalveolate lab; they might know.

> Bones will certainly help, and they are the only thing that
> can tell us 
> where to put the fossils. But analyzing them is so much
> work... The paper 
> we're talking about had pure calculation times of
> several months, but the 
> whole thing took maybe a year or two. The work by Livezey
> & Zusi took ten 
> years, didn't it, and it still contains way too few
> fossils...

Plus, too many bird fossils are fragmentary. If they had put them into their 
matrix, it would be a forest of "character state unknown"s. So I correct 
myself: we don't need just analyses of bones but new ways to analyze them. At 
least for birds. I checked out this 
http://www.fieldmuseum.org/research_collections/zoology/pdf/kearney_2002.pdf 
but it is of little help.


Eike


      __________________________________________________________
Gesendet von Yahoo! Mail.
Dem pfiffigeren Posteingang.
http://de.overview.mail.yahoo.com