Australopithecus and orangutans

[B.J.Tindall]

The question is not whether one accepts gene or protein sequence data, but
how one interprets that data and the possibility of congruence with other
data. I would like to draw your attention to the fact that in the paper you
quote the authors explicitly state in the abstract:

"analysis of the terminal sequences reconstructed the topology of the real
phylogeny and branch lengths accurately." Secondly they also assume a
bifurcating system

Most tree modelling systems assume bifurcation - I don't think we can
assume that this holds true in nature. Such a system also assumes that a)
the ancestors always become extinct and that b) we can always identify the
terminal taxa. If you were to analyse the genes of children, their parents
and the grand parents unless you knew in advance which were the "ancestors"
you would not get a "true reflection" of the "real phylogeny". In the case
of organisms with a known fossil record this may not be so much of a
problem, but the problem has become acute in the absence of a fossil record
i.e. in the case of prokaryotes.

The point is not whether one uses the sequence data, but what that data is
telling us. As a microbiologist I am fully aware of the contribution which
the sequence data has made to our understanding of prokaryote systematics,
but I am also aware of some of the problems which are there and are simply
ignored. The message is caution, as with most data sets.
Best wishes
Brian



At 11:36 15.5.2003 -0400, Jason de Koning wrote:
>>> What kind of argument is: " if molecular characters can never be used  
>>> for
>>> evolutionary analysis, then why do they so frequently give phylogenies
>>> that are congruent with "real cladistic analyses"? "
>>>
>>> For the same reason a stopped clock is correct twice a day is the  
>>> answer to that argument.
>>
>> That is a poor analogy, given the vastness of treespace.  For 15 taxa,  
>> there are ~2 x 10^14 possible rooted trees.  The probability of any  
>> given morphological and molecular datasets from the same 15 taxa to  
>> concur is thereby exceedingly and impossibly small, unless they  
>> contain mutual information (that is, historical signal).  I assumed  
>> that this would be a strawman argument for most people: an extreme  
>> projection of what is being argued is that molecular data contains no  
>> historical signal, which is preposterous.
>>
>> Since ultimately, the only way we have to really convince each other  
>> is through experimental or analytic tests of theoretical predictions,  
>> consider the several experimental phylogeny papers, where different  
>> groups have subjected bacterial strains or viruses to serial  
>> bottlenecks (or recently, generated mutants along a known phylogeny by  
>> PCR), sequenced genes from some of those strains, and then correctly  
>> reconstructed the historical process of simulated speciation.  See the  
>> following paper for a recent study:
>>
>> Sanson et al., 2002.  Experimental phylogeny of neutrally evolving DNA  
>> sequences generated by a bifurcate series of nested polymerase chain  
>> reactions.
>> http://www.ncbi.nlm.nih.gov:80/entrez/ 
>> query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11801745&dopt=Abstract
>>
>> If one chooses to ignore molecular data out of principle, one must  
>> confront their principles with evidence from in silico experiments  
>> (simulation), in vitro experiments (PCR), in vivo experiments  
>> (bacteriophage experiments), and from congruence between large  
>> morphological and molecular datasets, which all suggest that the  
>> cladistic analysis of molecular data (as defined conventionally) all  
>> lead to known, correct, or congruent answers.
>>
>> - Jason
>>
>


********************************************************************
* Dr.B.J.Tindall      E-MAIL bti at dsmz.de                           *          
* DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH *
* Mascheroder Weg 1b, D-38124 Braunschweig, Germany                *
* Tel.: ++ 531 2616 0 (general)                                    *
* Tel.: ++ 531 2616 224 (direct)                                   *
* Fax:  ++ 531 2616 418                                            *
*                                                                  *
* Homepage: http://www.dsmz.de/index.html                          *
* E-MAIL: contact at dsmz.de (general enquiries)                      *
*         sales at dsmz.de (sales)                                    *
********************************************************************