2 min read

One way to evaluate incomplete lineage sorting in your phylogenetic or phylogeographic data

As of this writing in 2011, there aren’t many ways to quantitatively evaluate incomplete lineage sorting (ILS) or ancestral polymorphism in molecular data in a statistically rigorous way.  One means of gauging relative amounts of lineage sorting is the gsi, which I will be working on later.  But, right here I wanted to note that I recently saw Ian Wang & Brad Shaffer using Wayne Maddison and Lacey Knowles method to evaluate ILS in a paper they wrote on dendrobatid frogs of Costa Rica, which is one of the main areas I am studying in my dissertation research.  Here is an excerpt from their paper:

“Because incomplete lineage sorting has the potential to confound
phylogenetic analyses of recently diverged populations, we
also performed the “minimize deep coalescences method” following
the procedure of Maddison and Knowles (2006). This method
takes into account the process of lineage sorting by genetic drift
and information on the genealogical relationships among alleles
(Maddison and Knowles 2006) and has been shown to improve the
probability of recovering an accurate population tree, even with a
single locus (Knowles and Carstens 2007). We first performed a
parsimony analysis in PAUP∗, using a heuristic search with 1000
random addition sequence replicates and maxtrees = 100, and
saved a majority-rule consensus tree from all equally parsimonious
trees recovered in the search. We then used the tree search
tool in Mesquite version 2.5 (Maddison and Maddison 2005) to
find the population tree that minimized the total number of deep
coalescences, using the saved majority-rule consensus tree and
SPR branch swapping (Maddison and Knowles 2006).”

Through this analysis, the authors were able to say that their phylogeny based on their mtDNA data was not different from a tree where ILS was minimized, suggesting the tree more likely reflected historical signal. 

This is a nifty little trick that I plan on using.  I am currently learning to implement gene tree modeling in MESQUITE, but I have a long way to go (feels like)….  my consolation is that I am already really familiar with the basics and with phylogenetic comparative analyses and trait evolution-related analyses in the program, so maybe my progress will be fast. 


Wang IJ, Shaffer HB (2008) Rapid color evolution in aposematic species: a phylogenetic analysis of color variation in the strikingly polymorphic strawberry poison-dart frog. Evolution, 62-11: 2742–2759.