Text 403, 70 rader
Skriven 2004-10-18 22:15:00 av Guy Hoelzer (1:278/230)
Ärende: Re: Can You Please Help M
=================================


in article ckjj1b$snf$1@darwin.ediacara.org, Michael Ragland at
ragland66@webtv.net wrote on 10/13/04 8:53 AM:

> "What is 'long-branch attarction' and how susceptible are the 3 major
> methods of phylogenetic reconstruction to it?"
> Thanks a bundle to anyone who can help.
> ---SuperNerdGirl
> 
> 
> 
> Cladistic parsimony(or maximum parsimony) is a method of phylogenetic
> inference in the construction of cladograms. Cladograms are branching
> tree like structures used to represent lines of descent based on one or
> more evolutionary change(s). Cladistic parsimony is used to support the
> hypothesis(es) that require the fewest evolutionary changes. It should
> be noted that for some types tree, it will consistantly produce the
> wrong results regardless of how much data is collected(This is called
> long-branch attraction).

This is true of long-branch attraction (LBA), but it doesn't really
distinguish it from other sorts of methodological biases.  Here is a little
more about it.  Branch length indicates the amount of independent evolution
that has caused the taxa at the ends of those branches to diverge from the
rest of the taxa included in the analysis.  When the extent of divergence
has become too large, the data for the taxa on the end of the branch
effectively becomes randomized relative to the data for the other taxa, so
it becomes less clear how the long-branch taxon fits into the hierarchical
pattern revealed for the rest of the taxa.  If you have only one taxon on a
relatively long branch, it will not be very clear where it should fit into
the tree.  If you have 2 taxa on relatively long branches, neither will fit
well with the rest.  No matter which method of tree topology estimation you
are using (assuming you are using one that is limited to extracting a
hierarchical tree topology), the 2 taxa appearing to have random data (data
that doesn't fit with the hierarchical pattern present for the other taxa in
the analysis) will tend to be joined as a clade because a pair of random
taxa will appear to share features relative to the set of taxa showing
hierarchical character-state (or distance) distributions.
 
> Using computer simulation to generate sequence data for protein
> families, we studied the influence of among site rate variation on the
> performance of phylogenetic reconstruction. The following methods for
> phylogenetic reconstruction were compared: maximum likelihood,
> parsimony, and distance-matrix analyses. Three different tree topologies
> were used to explore the effect of long branch attraction with data that
> incorporated substitution bias and different degrees of among site rate
> variation. 
> 
> While all methods showed errors due to long branch attraction under
> extreme conditions, maximum likelihood was by far the most successful in
> accurately resolving deep phylogenies

Use of computer simulations in this regard should be treated cautiously.  It
is very difficult to avoid assuming some (if not all) of the same aspects of
the evolutionary process in the generation of the data and in the likelihood
model assumed in the analysis.  This results in a circularity that can
easily mislead inferences in favor of the maximum likelihood method.  There
is no question that ML can be misled by LBA, like any other topology
estimation method constrained to yield pictures of bifurcating trees if the
assumed evolutionary model is incorrect, which is ALWAYS true in practice.

Guy
---
ū RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info@bbsworld.com

---
 * RIMEGate(tm)V10.2á˙* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 10/18/04 10:15:55 PM
 * Origin: MoonDog BBS, Brooklyn,NY, 718 692-2498, 1:278/230 (1:278/230)