[Taxacom] FW: Morphology vs Molecular

[Jason Mate]

> I find it interesting that it is recognized that some techniques failed. I am sorry if it seems that I am cherry picking the failed techniques. I am happy to acknowledge all failed techniques.
> And your acknowledgement of how one interprets phenetic analysis not being clear-cut does seem to pertain to my arguments. If its not clear cut then there is a problem.

My ackonowledgement pertains to the fact that our knowledge of all data is quite limited and this impacts our analysis. As such stating a priori that the character states of two taxa are a synapomorphy is just an opinion or educated guess.

> Yes, the key is comparison of homologous structures. In DNA sequences this seems to be determined by position, but recognizing the same position between species and higher taxa appears to have been problematic - and in my opinion it is one possibility by which erroneous similarity is retained or introduced in the analysis.

Same with anatomical characters. I.e. which digits were lost and which transformed in avian wings? Really, if you scratch deeper you begin to realise that ultimately, by your definition, most data is "probably, pretty-sure-it must-be, well I mean THEY LOOK THE SAME and are roughgly in the same area" homologous. To be a synapomorphy you not only need to compare the same but the change has to be the same (single origin). This level of knowledge is just not there and hence most synapomorphies are recognised a posteriori. In the more specific case of molecular data, different genes have different applications. Fast evolving ones are useless for ancient divergences either because of multiple hits erasing the evolutinary path or because they are so different (in particular due to intels) that establishing homology is difficult or impossible. Same issue with morphology.

> A undestand the principle that a synapomorphy is determined by the results of the analysis, but my particualr view is that the data being analyuzed is limited to those similarities that are shared derived for the group being analyzed. In morphology the derived condition can be articulated and justified for each indivudual character. In sequence analyses this appears to be more problematic - in my opinion.

Optimism can carry you far. An explanation of why a character in two or more species is homologous can be articulated with a similar degree of confidence regardless of the nature of the character (morphological or otherwise). Maybe the intagibility of sequences is disturbing but no more problematic as a source of data.

> More data may or may not help. In some case one might still end up with a 50:50 split between characters with 5, 100, or a million characters. 

It will not hurt. I can´t imagine the police not sending their CSI team to a crime scene in case it may not help. Try it and see.

> What is perhaps remarkable for the human-orangtuan example is that not only are there are as many morphological synapomorphies as found so far - at least 28, and may be at least 45, but that in molecular approaches this is not supported (interestingly, in some sequences I have scanned there are clusters for human-chimp, human-gorilla, human-African ape, but none or almost so for human-orangutan).

You have your pet peeve, I have mine. Like Maurice Chevalier said "so what, so what"

Best

Jason



check out the rest of the Windows Live™.
More than mail–Windows Live™ goes way beyond your inbox.
 More than messages
_________________________________________________________________
With Windows Live, you can organize, edit, and share your photos.
http://www.microsoft.com/middleeast/windows/windowslive/products/photo-gallery-edit.aspx