Posted by:
ScottThomson
at Fri Nov 5 20:25:09 2004 [ Report Abuse ] [ Email Message ] [ Show All Posts by ScottThomson ]
Not sure what you mean by "completely objective."
[/quote
Well basically that is me getting my back up against geneticists who claim that molecular analysis is the one, the only completelyobjective means to analyse phylogeny. But enough of that.
One of my biggest gripes against the way molecular analysis if often done is sampling presumed clades. They take a single specimen of a single species from a presumed species group and sequence one gene (and if its mtDNA its not really even that) and make a pronouncement on how revolutionary there findings on the relationships of the different clades are. Ridiculous, their analytical method does not meet any statistical challenge.
[quote]
You have also mentioned the lumping of species using allozyme and mtDNA data. I think that practice is untenable. Species consist of sets of actual or potentially interbreeding populations, and no mtDNA data can possibly tell us anything about reproductive isolation.
Yeah I probably should have put it the way Wulf did and say that the analysis failed to resolve the phylogeny. That was 3mtDNA genes and 1 Nuclear Gene by the way.
There is no doubt about it, the study of phylogeny is best done with molecules, simply because there are so many more molecular characters than morphological ones and because many molecules are selectively neutral whereas morphological characters tend to be adaptive and therefore more likely to be the result of convergent evolution. Nevertheless, phylogeny per se really gives us an incomplete picture.
Really, I have seen no evidence of that. Many molecular based trees are outright ridiculous, others are very good. Some tell us very important things leading to valid reassessment of the phylogeny including the morphology. There is no evidence that a given set of molecular sequences are neutral, they could be adaptive, neutral or random. Knowing which is which is like throwing darts at a dart board with a blindfold on. The molecular trees hope that enough of the molecules are neutral to override those that are deliterous to the tree. There is no way to know this, it is an assumption and this is why Molecular data is based on subjective analysis.
There also seems to be still a general methodological problem in molecular systematics. The weighting of data as preferred by some taxonomists may influence phylogenetic relationships. "Subjective" weighting or rejecting of data seems to be more like a "...if they woun't fit, I'll make them fit..." thing and there are several methods to handle data sets used in phylogenetic studies.
I agree, as I have outlined above. There seems to be a definite belief that they are outside the boundaries of statistically valid evidence. I harp on that one, yes, but I did minor in statistics. There are huge assumptions in the weighing of data because there is not necessarily any idea of what the genes in question do, so how can you weigh them. I do on occasion weigh morphological data but at least I can use the function of the character to guide the weighing. It too is not perfect and contains assumptions but I test this by getting the identical tree with weighed and unweighed data, the only differences is nodal support.
I use molecular data extensively, it is not all bad, but it is one tool. That is all, there are others which include morphology, and statistical analysis. Morphology is effected, visibly, by behaviour, so understanding the ecology and ethology of the species is also important. Good phylogenies require the use of as many tools as possible. Not one.
Cheers, Scott
Carettochelys.com
-----
Scott Thomson
If you believe you can or you can't you are always right.
[ Reply To This Message ] [ Subscribe to this Thread ] [ Show Entire Thread ]
|
