46. Felsenstein, J. 1985. Phylogenies and the Comparative Method. American
Naturalist 125:1-15.
Blog Author:Ben Clinch
Blurb Author: T. Jonathan Davies
-Department of Biology, McGill University
-Interest in macroecology of infectious disease
-Couldn’t find much on this guy
Paper Author: Joseph Felsenstein
-Professor, Departments of Genome Sciences and Biology, U. of Washington
-Best known for work on phylogenetic inference, author of Inferring Phylogenies
-Introduced methods for making statistically independent comparisons using phylogenies
-Also noted for his work in theoretical population genetics- selection, migration, linkage, speciation, and the coalescent
-Member of the National Academy of Sciences
-Awarded Darwin-Wallace Medal in 2008, John J. Carty Award in 2009, International Prize for Biology in 2013
-Moth species Ufeus falsensteininamed in his honor
Main Question: Can phylogenies be corrected for when using comparative methods?
Background:Prior to the proposed comparative method in this piece, traditional comparative methods involved correlating phenotypes and environment for a species range. These traditional methods treated species as independent, even though they are inherently non-independent due to the clustered structure of phylogenies.
Assumptions:
-Phylogenetic inertia is absent
-Characters in a phylogeny evolve by Brownian motion that is independent in each lineage
Methods:
-Use Brownian model to obtain seven independent contrasts of both X and Y phenotypes
-X contrasts will have covariance with Y contrasts of the same points
-4 step process for computing contrasts
-Can test for independence of evolution between X and Y by simply testing whether the correlation is zero
Results:
-Proposed model with ability to compute independent contrasts from any phylogeny.
-Proposed method to correct for phylogeny requires tree topology knowledge and branch lengths, as well as the use of Brownian motion modeling on a linear scale.
Conclusions:
-This piece is a good step, but there still isn’t an acceptable method of correcting for the presence of phylogenies
-Difficulties:
-How to reconstruct phylogenies?
-How to put confidence intervals on inferences?
-What if we lack an acceptable statistical model of character change?
-What if we don’t take the phylogeny into consideration?
-Most regression, correlation, and contingency methods take points to be drawn independently. Phylogenies are structurally non-independent, so using methods that assume independence will overstate significance in tests.
Comments/Questions:
-The blurb mentions the constraints which held the method’s full potential back. Have recent advances allowed for better use of Falsenstein’s model?
-What is the parsimony method used by Gittleman and Ridley? There’s a lot of detail about its problems but not really what it is exactly.
-This wasn’t the easiest paper to digest, there was a lot to chew on but makes a bit of sense after re-reading.
I enjoyed reading this paper. I thought it was relatively easy to follow and that his points were logical. As someone who has used phylogenetic techniques in the past, I appreciated getting to read the first paper that explains why it is important to consider phylogenetic relationships when doing comparisons.
ReplyDeleteI agree with Maria, it was good to read the first paper first to understand the importance of phylogenies and understand that that was an issue with past models. I think this paper was very important for the time and really laid the ground work for future research in this area, even addressing the aspects that are flawed.
ReplyDeleteI like this paper a lot. It's pretty easy to follow. It really opens up my eyes about how phylogenetic trees are constructed.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteI agree, it was fun to read this paper. I think that with his hypothetical examples, Felsestein makes a very good job at demonstrating why we need to care about phylogeny in comparative analyses.
ReplyDeleteIn order to understand species relationships, phylogeny is the way to go but the interpretation of each speciation and their character mapping is pretty hard. I have read Felsenstein's several papers before and he is one of the early champions of phylogenetic research. This paper is easy to read and explains the fundamental phylogenetic approach its statistical component of the analysis. I am sure this paper was important at that time to improve the technique. After the genetic information is available for most of the species people tend like molecular phylogeny over morphological comparative analysis which I don't agree. Comparative morphological phylogeny is still important and valid. It is really hard to show hybridization events but someone will figure this out in phylogenetic studies.
ReplyDeleteI loved reading this paper. It had a really nice break down and this helped big time with aiding in my understanding of the paper. In figure 5, what do they mean by a "worse case"? Is that a worse case scenario as of what the graph can look like, what could possible happen, or what?
ReplyDelete