Reality

[B.J.Tindall]

Richard,
Interesting reply and thanks for the link. One point you also raised was
the fact that molecular data contain "more data" than morphologial data.
Well yes and no.
a) to understand orthology of a gene you have to know what it is doing,
i.e. what it codes for, and you very quickly come back to the structure and
function of the gene product (usually proteins) - "molecular morphology".
b) If you take something like the 16S rDNA sequence in prokaryotes you have
about 1400-1500 bases. However if the two species you are comparing share
90% sequence similarity you will only find 140-150 bases "have any say" in
the matter because the rest are identical. If you then assume that 1-2% of
the gene sequence is "hyper-variable" and you can't rely on it, then you
are down to about 120-130 characters. Okay this may be more than a
morpholological data set can offer, but at 90% ribosomal gene sequence
similarity in eukaryotes you are dealing with a plant and an animal species
anyway (correct me if I am wrong). So when you compare animal species you
may be dealing with a greater degree of seqeunce similarity, and even less
bases "have anything to say". It is not my intention to reduce the
significance of the molecular data, but one has to be a bit careful,
numbers have been thrown around in the past and they have been proven to be
"biologically misleading".
Brian

At 10:12 16.1.2002 -0500, Richard Zander wrote:
>Well, let's see. We have equal evidence supporting two different
hypotheses: 3 gene trees supporting, say, ((AB)C) and 3 gene trees
supporting ((AC)B). I figure you have a 50/50 chance of guessing the
correct tree.
>
>If you have 0 contradictory gene trees, you need only 3 agreeing gene
trees to match or exceed 95% confidence level that you have demonstrated a
species tree. If you have 1 contradictory tree you need at least 5 agreeing
trees. With 2 contradictory trees, you need 7 agreeing trees.
>
>More on this at:
>http://www.buffalomuseumofscience.org/botany/num/num_trees.htm
>
>Yes, it is not easy to find decent molecular evidence of a probabilistic
phylogenetic relationship through falsifying (at some confidence level) the
null that the amount of support for each of the three arrangements of
terminating branches on an internal branch is randomly generated and equal.
It is possible, though.
>
>---------
>From:
>Richard H. Zander
>Emeritus Curator of Botany
>Clinton Herbarium
>FNA Editorial Center at Buffalo
>
>The Buffalo Museum of Science
>1020 Humboldt Pkwy
>Buffalo, NY 14211 USA
>
>email: rhzander at sciencebuff.org NEW EMAIL ADDRESS! (Note "h" infix)
>voice: 716-896-5200 x 351   FAX: 716-897-6723
>home: http://www.buffalomuseumofscience.org/div_botany.html
>BFNA home: http://www.buffalomuseuofscience/BFNA/BFNAmenu.htm
>FNA home: http://hua.huh.harvard.edu/FNA/
>
>
>----- Original Message -----
>From: "B.J.Tindall" <bti at DSMZ.DE>
>To: <TAXACOM at USOBI.ORG>
>Sent: Wednesday, January 16, 2002 9:54 AM
>Subject: Re: Paraphytlly is real.
>
>
>> Okay, but this assumes that the data set conforms to the assumption laid
>> down in bootstrapping before I do the test. One of the problems is
>> independence in the data set - do adjacent bases in a gene sequence evolve
>> independant of each other?
>> Perhaps you could answer a simple question - you have three gene trees
>> which agree and then you have a set of 3 other genes which agree with
>> eachother but disagree with the first set - so which one is correct?
>> Given the rate at which microbial genomes are being sequenced the first
>> indications suggest it may not be that easy to find 3 genes which agree in
>> detail. Perhaps we have to live with that fact?
>> Brian
>


********************************************************************
* 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                                            *
* Fax:  ++ 531 2616 491 (ISDN)                                     *
*                                                                  *
* Homepage: http://www.dsmz.de/index.html                          *
* E-MAIL: help at dsmz.de (general enquiries)                         *
*         sales at dsmz.de (sales)                                    *
********************************************************************